Interactive effects of drought and shade on three arid zone chenopod shrubs with contrasting distributions in relation to tree canopies


  • J. N. Prider,

    Corresponding author
    1. Discipline of Environmental Biology, School of Earth and Environmental Sciences, The University of Adelaide, Adelaide, South Australia 5005, Australia
      †Author to whom correspondence should be addressed. E-mail:
    Search for more papers by this author
  • J. M. Facelli

    1. Discipline of Environmental Biology, School of Earth and Environmental Sciences, The University of Adelaide, Adelaide, South Australia 5005, Australia
    Search for more papers by this author

†Author to whom correspondence should be addressed. E-mail:


  • 1Plants that grow beneath trees in arid systems may frequently experience both water and light limitation, although protection from high radiation loads during drought may compensate for a loss of productivity due to reduced light availability when water is plentiful.
  • 2We examined the effects of shading, during an imposed water deficit, on the carbon gain, stomatal conductance (gs) and shoot water potential (Ψs) of seedlings of three shrubs: Atriplex vesicaria (Heward ex Benth.), a C4 species, and Enchylaena tomentosa (R. Br.) and Rhagodia spinescens (R. Br.), which are restricted to shaded sites beneath trees.
  • 3Under conditions of limiting water, photosynthetic rates measured at saturating light (Amax) were negative in high-light grown Enchylaena plants but remained positive in shade-grown plants. When water was not limiting, Amax was reduced in shade-grown Atriplex but shade did not affect carbon gain in the other two species.
  • 4AtriplexΨs was higher in shaded than in unshaded plants, but in unshaded plants positive carbon gain was maintained at Ψs below −10 MPa. Stomatal conductance and Amax decreased more slowly with increasing water deficit in shaded conditions in all species.
  • 5Atriplex was tolerant of a broader range of light and soil moisture conditions than Enchylaena, with Rhagodia intermediate between these two species. The interactive effect between drought and shade and the ecophysiological tolerances of these three species have consequences for their field distributions.


While plant growth in arid systems is strongly limited by water availability, nutrients and light may also limit growth when water is available. In many semiarid to arid systems the relative abundances of these resources are modified beneath plant canopies (see reviews by Hunter & Aarsen 1988; Belsky et al. 1989; Callaway 1995). Although light is reduced in these microhabitats, nutrient and soil moisture availability may be higher, and evaporative demands lower, due to increased humidity and/or reduced temperatures. Such modifications may enhance the growth of plants in this microhabitat (Callaway 1995). Whether a plant growing beneath the canopy of another is facilitated depends on its physiological tolerances and minimal resource requirements, especially for light and soil moisture at different life stages (Holmgren, Scheffer & Huston 1997). Seedlings are particularly sensitive to resource shortages and stresses associated with high temperatures and radiation loads, and low moisture availability.

Researchers have examined the combined effect of shortages of light and soil moisture on the growth of temperate and tropical woodland and forest seedlings (Fisher, Howe & Wright 1991; Veenendaal et al. 1996; Baruch, Pattison & Goldstein 2000; Poorter & Hayashida 2000; and review by Coomes & Grubb 2000). Many of these experiments have compared species that typically establish in canopy gaps with those that can establish in the deeply shaded understorey. Some researchers have found that plants growing in high-light conditions have greater sensitivity to soil moisture status; water limits growth when light is not limiting, but light limits growth when water is not limiting (Kozlowski 1949; Fisher et al. 1991; Muraoka et al. 1997; Baruch et al. 2000; Holmgren 2000; Poorter & Hayashida 2000). However, the light climate beneath canopies in arid systems differs from that in many temperate or tropical systems. As rainfall decreases or the length of the dry season increases, light penetration through the canopy increases (Coomes & Grubb 2000). The shade cast by shrubs and trees in arid systems is not likely to be dense, but the effects of light reduction on the carbon gain of understorey plants in these systems have not received much attention. Whereas shade may be detrimental to seedling growth in mesic systems, shade may have a positive effect on seedlings in arid systems (Holmgren et al. 1997). In arid systems, severe shortages of soil water often coincide with periods of high temperatures and high solar radiation, producing multiple stresses on plant performance. Protection from high radiation loads in shaded microenvironments during drought may compensate for a loss of productivity due to reduced irradiance when water is available.

Smith & Huston (1989) predicted for forest ecosystems that there was a trade-off between shade tolerance and drought tolerance. Plants adapted to shaded conditions would be more sensitive to drought than plants adapted to higher irradiation because of biomass allocation patterns under dry and low-light conditions. Plants growing in low light allocate a greater proportion of biomass to light-capturing organs, creating a larger transpirational area, and are thus susceptible to dry conditions. But under dry conditions, proportionally more biomass is allocated to roots which increases the ratio of nonphotosynthetic tissue, hence the requirement for more light to maintain a positive carbon balance. Consequently, it has been hypothesized that growth and survival in shaded microenvironments become possible only with increased moisture availability in these sites (Holmgren et al. 1997). However, plants may be able to persist in shaded habitats in arid systems when reduced evaporative demand below canopies reduces water stress for plants growing there. Shading may therefore lessen the impact of drought by reducing water loss from soils and plants (Holmgren 2000; Hastwell 2001). Alternatively, light and soil moisture may have independent effects, such that the effects of drought on plant performance may change proportionally with decreasing light availability (Sack & Grubb 2002).

In arid open woodlands in southern Australia, Acacia papyrocarpa (Benth.) trees provide microhabitats beneath their canopies with modified temperature and light climate. The higher organic matter content of soils beneath A. papyrocarpa canopies (Facelli & Brock 2000) improves soil macroporosity and texture, which can affect the infiltration and retention of water (Cowling 1978; Joffre & Rambal 1988) and root penetration (Greacen & Oh 1972; Pugnaire et al. 1996). These modifications to soils and microclimates have beneficial effects on the growth and survival of some plants occurring in these microhabitats (Prider 2002). Two shrubs, Enchylaena tomentosa (R. Br.) and Rhagodia spinescens (R. Br.), occur almost exclusively beneath tree canopies in these woodlands (Facelli & Brock 2000). In contrast, one of the commonest chenopod shrubs in the system, Atriplex vesicaria (Heward ex Benth.), has a distribution unrelated to the presence of canopies (Tester et al. 1987; Facelli & Brock 2000). Pot-grown seedlings of these three shrub species were used to compare their responses to an imposed water deficit when grown under irradiance simulating the light conditions beneath A. papyrocarpa canopies and the open sites between canopies. The study addressed the following questions: (1) Does shade affect physiological responses to water deficit (shoot water potential, stomatal conductance, carbon gain)? (2) Is the carbon gain of plants typically found only beneath canopies (E. tomentosa and R. spinescens) sensitive to high light, particularly during drought? (3) Is the carbon gain of plants found in high-light environments (A. vesicaria) affected by shade?

Materials and methods


Atriplex vesicaria (Chenopodiaceae) is a small, upright, perennial shrub to 1 m high, common throughout a variety of habitats in southern arid and semiarid Australia. Rhagodia spinescens (Chenopodiaceae) is a sprawling perennial shrub to 1·5 m tall. It is frequently found in shaded habitats or along perennial watercourses in the drier parts of its range. Enchylaena tomentosa (Chenopodiaceae) is a small, perennial shrub to 1 m high, and is also commonly found in shaded areas within the drier limits of its range. All three species occur within the A. papyrocarpa woodlands at Middleback Field Centre, South Australia (32°57′ S, 137°24′ E). This region experiences an annual rainfall of 217 mm (75 years average) with an average daily maximum temperature during January of 28·9 °C (Facelli & Brock 2000). Soils in the region are structureless red calcareous earths of clay-loam texture (Jessup & Wright 1971). Soils under mature A. papyrocarpa canopies have higher nutrient levels, and lower water retention and bulk density, than soils beyond the canopy edge (Facelli & Brock 2000). All species are subsequently referred to by generic name.

experimental design

Seedlings of Atriplex, Rhagodia and Enchylaena were grown in factorial combinations of two irradiances and two soil types. Seedlings were raised from purchased seed sourced from populations in areas with <400 mm annual rainfall (Blackwood Seeds, Murray Bridge, South Australia). Seeds were sown in 0·3 l seedling tubes filled with soil collected from under Acacia canopies (canopy soil) or from adjacent open spaces (open soil). Both soil types were used, as soils differed in their water-holding and water-retention characteristics, which could affect water availability during the drought period. The field capacity of canopy soils was 26% soil dry weight (DW); and that of open soils, 22% soil DW. Plants did not recover turgor when rewatered after the water content of canopy soils dropped to 5·6%, or that of open soils fell to 5%. Each plant was fertilized with 1·4 g (recommended rate) Osmocote slow-release fertilizer (N : P : K 17 : 1·6 : 8·7; Scotts Australia Pty Ltd) to eliminate the potential confounding effect of different nutrient levels in the two soil types. A total of 200 seedlings of each species were grown, giving 50 replicates for each treatment of soil type and light. Pots were kept outdoors, packed into polystyrene foam boxes in random order, and mulched with a 20 mm layer of vermiculite to reduce evaporative losses from the soil surface and lower soil temperatures. Shade was provided to half the pots by black shade cloth (Weathashade), which has a neutral effect on light quality (Yates 1989). The incident photosynthetically active radiation (PAR) under shade cloth and Acacia trees at the field site was measured with light sensors (Li-Cor LI-190SZ Quantum sensors; Li-Cor Inc., Lincoln, NE, USA). The shade cloth reduced incoming PAR to 20% of ambient unshaded conditions, which represents a level of shading similar to that produced by Acacia trees at the field site. The remaining plants were grown in full sunlight. Seedlings were watered every other day unless the temperature was >30 °C, in which case they were watered daily. Only the surface soil of unshaded pots dried out between waterings. Plants were hardened by omitting water for 3 days on two occasions prior to the imposition of drought conditions.

physiological measurements

Plants were grown in pots for 4 months from the time of seed sowing. Measurements of shoot water potential, stomatal conductance and carbon gain commenced on the last day the plants were watered. Measurements were then made daily (two replicates per treatment) for 10 days on unwatered plants until carbon gain was no longer positive. As the aim of the experiment was to examine plant responses to summer drought, the drought period occurred during February 2000 (mean air temperature ± 1 SD, 36·2 °C ± 3·5). Photosynthetic rates at saturating light and stomatal conductance were measured on terminal shoots of plants using a Li-6200 Portable Photosynthesis System (Li-Cor). Before measurements were made between 10 am and 12 pm, plants were left to equilibrate for 30 min to the light conditions under which the measurements were to be made. After measurement, each shoot was excised and immediately inserted into a Scholander-type pressure chamber for measurement of shoot water potential. Shoot water potential measurements represent plant water status at the time carbon assimilation measurements occurred, and not at equilibrium with soil water status. The leaf area of measured shoots was determined by scanning excised leaves and calculating scanned leaf area with an optimas imaging program ver. 5·2 (Media Cybernetics Inc., Silver Spring, MD, USA).

The remaining above-ground biomass was cut at soil level and dried to constant weight at 85 °C. Shoots on which leaf area had been determined were dried and weighed separately for determination of specific leaf area. Specific leaf area was calculated as A/M (cm2 g−1), where A is leaf area and M is leaf DW. To determine the soil water content (SWC) of each pot at the time when physiological measurements were made, pots were weighed for wet soil weight immediately after each measurement. Dessication of soil within pots was prevented until the roots could be extracted, usually within 5 days. The fresh and dried weight of roots was determined. Soil from the pots was oven-dried to constant weight at 105 °C. The SWC in each pot was determined gravimetrically after first subtracting the weight of fresh roots from the initial wet soil weight.

statistical analyses

Shoot water potential (Ψs), stomatal conductance (gs) and carbon assimilation rate at saturating light (Amax), were dependent variables in separate analyses of covariance (jmp in ver. 4·0.3, SAS Institute, Cary, NC, USA) with percentage SWC as the covariate. All physiological variables were log-transformed for analyses to satisfy assumptions of linearity. Significant differences in slopes in these models (interactions between the covariate and other effects in ancova models) indicated different responses of the physiological variables over the drought period for different experimental treatments. Other significant effects in ancova models, which did not include the covariate, indicated significant differences in the elevations or intercepts of the regression lines. In order to test differences between treatments at both low and high SWC, values of dependent variables were calculated at two SWC levels, 8 and 20%. These soil water contents were selected as physiological variables were measured across all treatments within this range of soil moisture.


drought period

Over the 9–10-day drought period, the SWC in both soil types fell from 24·2 ± 1·0% (mean ± 1 SE) of soil DW across treatments on the final day of watering to 8·4 ± 0·6% of soil DW (Fig. 1). Over time, shaded soils maintained higher SWC than unshaded soils; however, this does not affect the conclusions because physiological measurements were analysed as responses to SWC rather than time since last watered. Atriplex maintained positive carbon assimilation for 10 days after cessation of watering, but carbon assimilation rates had become negative in Enchylaena and Rhagodia after 4 days in unshaded conditions and after 8 days in shaded conditions. Li-Cor chamber temperatures during measurements were significantly higher in unshaded than in shaded conditions (means ± 1 SD, 38·7 ± 2·9 °C and 34·8 ± 3·0 °C, respectively). However, there was no significant relationship between temperature and photosynthetic rate over the measurement period and across the range of temperatures recorded (r2 = 0·003, P = 0·45).

Figure 1.

Soil water content (SWC) as a percentage of soil dry weight for shaded (▴) and unshaded (□) treatments over the drought period.

shoot water potentials)

All factors – species, light and soil type – had a significant effect on the relationship between Ψs and SWC (Table 1). With the exception of Atriplex plants grown in shade, Ψs of all plants declined rapidly over the drought period (Fig. 2). The slope of the relationship between Ψs and SWC varied with species and light intensity (species × light × SWC interaction, Table 1). For example, Atriplex and RhagodiaΨs decreased at a faster rate under unshaded than under shaded conditions (Fig. 2). Shade-grown Atriplex plants had the slowest rate of decline in Ψs, with the fastest occurring in unshaded Rhagodia plants (slope comparisons, Fig. 2). In contrast, the Ψs of well watered plants did not differ with species or light treatments.

Table 1.  Results of ancova of Ψs, gs and Amax with soil water content (SWC) as the covariate in each case; testing the effects of species, light and soil type and their interactions
SSF ratioSSF ratioSSF ratio
  • All values were log-transformed before analysis.

  • Significant effects that include the covariate indicate different slopes; see Fig. 1

  • and Fig. 2 for multiple comparisons of slopes.

  • Significant F values are labelled at the following probability values: *α < 0·05, **α < 0·01, ***α < 0·001.

SWC (covariate)146·89533·61***95·68268·49***10·02186·97***
Species2 0·57  3·24* 1·15  1·61 0·24  2·24
Soil1 0·52  5·91* 1·91  5·36* 0·50  9·36**
Light1 0·81  9·19** 1·38  3·87* 0·04  0·73
Species × soil2 0·74  4·22* 4·48  6·29** 0·18  1·72
Species × light2 1·21  6·89** 5·17  7·26** 0·38  3·54*
Soil × light1 0·46  5·26* 0·65  2·12 0·26  4·80*
Species × soil × light2 0·04  0·21 1·07  1·50 0·01  0·08
Species × SWC2 1·14  6·46** 6·39  8·97** 0·04  0·37
Soil × SWC1 0·23  2·67 1·85  5·20* 0·01  0·15
Light × SWC1 1·10 12·50*** 3·33  9·36** 0·50  9·33**
Species × light × SWC2 0·82  4·68* 0·76  1·07 0·05  0·43
Species × soil × SWC2 0·22  1·25 2·93  4·10* 0·05  0·45
Soil × light × SWC1 0·02  0·23 0·03  0·07 0·02  0·36
Soil × light × species × SWC2 0·29  1·66 0·10  0·15 0·03  0·31
Figure 2.

Changes in shoot water potential (Ψs) with increasing soil water content (SWC) for Atriplex, Enchylaena and Rhagodia grown in shaded (▪) and unshaded (□) conditions. Left-hand figures show raw data with curve fits to illustrate response shapes (two-site binding hyperbolas, graphpad prism ver. 3·02); right-hand figures show regressions between log-transformed variables (x-axis, 2 = low SWC, 3 = high SWC). The values in these plots are the slopes of each regression (n, no shade; s, shade). Values labelled with a different letter were significantly different at α < 0·05 (Tukey-type tests, Zar 1999). *, Significant difference between light treatments, within species, at either low or high SWC (Tukey-type tests, Zar 1999, α < 0·05).

Soil type significantly affected the Ψs of well watered plants growing at high light, Ψs being higher in open than canopy soils, but soil type had no significant effect on Ψs of droughted plants in either light treatment (soil × light interaction, Table 1). There was also a species × soil type interaction (Table 1). EnchylaenaΨs was higher in droughted canopy soils, and AtriplexΨs was higher in well watered, open soils.

stomatal conductance (gs)

As SWC decreased, stomatal conductance fell at a slower rate in shaded than in unshaded treatments (light × SWC interaction, Table 1, slopes 1·96 and 1·45, respectively). The effect of light on gs differed between species subject to drought (species × light interaction, Table 1; Fig. 3). Stomatal conductance was significantly lower in shaded Atriplex and Enchylaena plants at both high and low SWC, but Rhagodia gs was lower only in shaded plants at high SWC (Fig. 3).

Figure 3.

Changes in stomatal conductance (gs) with increasing soil water content (SWC) for Atriplex, Enchylaena and Rhagodia grown in unshaded (□) and shaded (▪) conditions. For details see caption to Fig. 2.

The relationship between SWC and gs differed between species in different soil types (species × soil × SWC interaction, Table 1). In Enchylaena and Rhagodia, gs fell at a slower rate in canopy soils than open soils. In Atriplex, gs fell at significantly slower rates in drying open soils than in canopy soils. Stomatal conductance differed significantly between soil types at high and low SWC, but this was not consistent for all species (species × soil interaction, Table 1). In well watered soils, Rhagodia and Enchylaena gs was higher in open soils than canopy soils. For Atriplex gs, there was no difference between soil types at high SWC, but at low SWC gs was higher in open soils than canopy soils.

photosynthetic rate (amax)

Amax fell at a slower rate over the drought period for plants growing in the shade (light × SWC interaction, Table 1, slopes 0·74 and 0·48, respectively). There were species differences in Amax between light treatments at low and high SWC (species × light interaction, Table 1). At high SWC, only Atriplex plants were significantly affected by the light treatment under which they were grown, with Amax lower in shade-grown plants (Fig. 4). At low SWC, only Enchylaena plants were significantly affected by light. High-light-grown plants had stopped gaining carbon at the time of measurement, while plants growing in the shade maintained carbon uptake at this time (Fig. 4).

Figure 4.

Changes in photosynthetic rate at light saturation (Amax) with increasing soil water content (SWC) for Atriplex, Enchylaena and Rhagodia grown in unshaded (□) and shaded (▪) conditions. For details see caption to Fig. 2.

Amax was higher in Atriplex than in other species over a wide range of gs (Fig. 5). The relationship between gs and Amax was strongest for Rhagodia (r2 = 0·70, on transformed values), but changes in gs explained only 48% (r2 = 0·48) of the variation in Amax for Atriplex and 56% for Enchylaena. There were also differences between light treatments in the strength of this relationship. For Atriplex, gs explained 68% of the variation in Amax in unshaded treatments but 38% in shade treatments; for Enchylaena, 83 and 33%; and for Rhagodia, 78 and 63% in unshaded and shaded treatments, respectively.

Figure 5.

Relationship between gs and Amax for Atriplex (□), Enchylaena (▵) and Rhagodia (○). Curve fits are to illustrate response shapes (two-site binding hyperbolas, graphpad prism ver. 3·02).

Amax differed between soil types in the two light treatments, but the differences between soil types were significant only at low SWC (soil × light interaction, Table 1). In dry soils, carbon gain was negative in plants growing in canopy soil in unshaded conditions, but there were no differences between other treatments, which all had positive carbon gain.


Shoot and root biomass differed between species, soil types and light treatments, but there were interactions between pairs of these factors (Table 2). All species had less root biomass in shaded than in unshaded treatments, yet only Atriplex shoot biomass was reduced when grown in shade (Fig. 6a). Rhagodia had less shoot biomass than the other two species, and less root biomass than the other species under unshaded conditions (Fig. 6a). Shoot biomass was also affected by soil type. Rhagodia shoot biomass was significantly greater in canopy soils than open soils, and there was a similar trend for Enchylaena (Fig. 6b). Soil type also interacted with light; shoot and root biomass was greater in canopy soils than open soils, but only in unshaded treatments (Fig. 6c).

Table 2.  Results of separate anova testing the effects of species, light and soil type on shoot, root and total biomass
Effectd.f.Shoot biomassRoot biomassTotal biomass
SSF ratioSSF ratioSSF ratio
  1. Significant F values are labelled according to the following probability values: * α < 0·05, ** α < 0·01, *** α < 0·001, #P = 0·053.

Species29·7313·52***0·53 22·59***11·5215·85***
Soil18·1522·66***0·22 18·94*** 7·1819·75***
Species × soil23·67 5·10**0·04  1·90 3·75 5·16**
Species × light26·08 8·45***0·30 12·83*** 5·91 8·14***
Soil × light11·62 4·51*0·16 13·60*** 1·38 3·80#
Species × soil × light20·47 0·650·04  1·82 0·43 0·59
Figure 6.

Mean (+ 1 SE) biomass of Atriplex (A), Enchylaena (E) and Rhagodia (R) plants grown in shaded (S) or unshaded (N) conditions and canopy (C) or open (P) soils, showing (a) species × light; (b) species × soil; and (c) soil × light interactions. Shaded parts of columns indicate root biomass; unshaded parts, shoot biomass. Column parts labelled with different letters indicate significant differences between either root or shoot biomass for each treatment combination where interactions were significant (Tukey–Kramer HSD tests, α < 0·05).

Shoot biomass as a proportion of total biomass was similar for Atriplex and Enchylaena, but Rhagodia had a significantly greater proportion of shoot biomass (anova, F = 9·048, P < 0·001). Shoot proportions and specific leaf area were significantly greater in shaded than in unshaded treatments (anova, F = 253·300, P < 0·001; F = 28·284, P < 0·001, respectively).


The overall effects of shade in arid environments may be positive when water becomes limiting. Plants had slower declines in Ψs (with the exception of Enchylaena), gs and Amax with increasing water deficit when grown in shade than when grown in full sun. This suggests water stress is mitigated by shade during the summer months, and that plants in shaded microclimates can maintain a positive carbon balance for longer into drought periods than the same species in unshaded microclimates. However, the three species responded differently to water stress in the contrasting light conditions. Differential responses were expected between species naturally occurring in different light environments, but even the two species found in shaded environments, Rhagodia and Enchylaena, had varied responses. Their shade dependence may be due to different combinations of physiological traits.

water relations

Slower declines in gs in shaded plants, and high gs in unshaded, well watered plants, suggest that evaporative demand under shade was not as great as under high irradiance when ambient air temperatures and vapour pressure deficits are high. Because of lower temperatures in shade, shaded leaves require less water for transpirational cooling, and plants are able to conserve more water and maintain higher tissue water status (Schultz & Matthews 1997; Valladares & Pearcy 1997; Sawada et al. 1999). Atriplex responded differently from the other two species to changes in evaporative demand under shaded conditions by maintaining high Ψs in shade as SWC fell. One possible explanation is changes in stem architecture. Unlike Rhagodia and Enchylaena, shade-grown Atriplex plants had very thin, lax stems as compared to the thicker, upright stems of plants grown in high light. Presumably xylem development (either the number and size of vessels) in shaded plants was not as extensive as in high-light-grown plants. For plants with poor hydraulic architecture, maintaining high tissue water status by decreasing water potential gradients may reduce the risks of cavitation (Schultz & Matthews 1997), which could increase stem survivorship even though there may be reductions in growth.

Although shade-grown plants had higher proportions of shoot biomass than high-light-grown plants, transpirational water losses per leaf area may not be greater in shade-grown plants, contrary to the predictions of Smith & Huston (1989). Reduced xylem formation (Schultz & Matthews 1993) and leaf vascularization (Kozlowski 1949), and fewer stomata (Osmond, Björkman & Anderson 1980; Valladares & Pearcy 1997), may reduce transpirational water losses due to anatomical differences between high- and low-light-grown leaves. Coupled with reductions in water and heat stress in the shade, shade becomes an advantage as conditions become drier, as it improves plant water relations (Amundson, Ali & Belsky 1995; Dalton & Messina 1995; Holmgren 2000; Shumway 2000; Maestre et al. 2001; Kröpfl et al. 2002; but cf Valladares & Pearcy 2002). Such effects may have enabled shaded Enchylaena and Rhagodia to continue gaining carbon longer into the drought period. However, the Ψs of these shade-dependent species responded differently to increasing water deficit. Whereas the Ψs of Enchylaena was unaffected by light conditions as water deficit increased, RhagodiaΨs declined rapidly in full sun, suggesting that this species may use more water when growing in these conditions.

carbon gain

In high-light treatments, Atriplex maintained higher photosynthetic rates over a broader range of Ψs than Enchylaena or Rhagodia. Atriplex also had a higher leaf-level photosynthetic rate for a given gs than the other two species, and carbon gain was only equal between species at very low SWC. Thus Atriplex had a more conservative water-use pattern or water-use efficiency, being able to fix more carbon with less transpirational water loss. Atriplex vesicaria utilizes the C4 carboxylation pathway, while Enchylaena and Rhagodia are C3 plants. Although the difference in water-use efficiency between C3 and C4 species can be greater at high than at low soil water contents, C4 plants use water more conservatively so they are able to remain physiologically active for longer into drought periods (Chapman & Jacobs 1979; Kalapos, van den Boogaard & Lambers 1996). In this study, Atriplex was able to gain carbon for longer into the drought period, whereas photosynthesis ceased in high-light-grown Enchylaena plants soon after the imposition of drought. Rhagodia was able to maintain positive carbon gain to lower Ψs than Enchylaena, but as in Enchylaena, carbon gain ceased soon after watering stopped.

Decreases in Amax in unshaded plants over the drought period were influenced more by stomatal closure than in shaded plants, at least for Atriplex and Enchylaena. Similar patterns have been found in other comparisons of shaded and unshaded plants (Lipscomb & Nilsen 1990a; Lipscomb & Nilsen 1990b; Niinemets et al. 1999; Sawada et al. 1999). Non-stomatal limitations to carbon assimilation are still poorly understood (Nilsen & Orcutt 1996), although it has been suggested that damage to both carbon reduction and light-harvesting functions in response to water stress may limit carbon assimilation under water stress (Gamon & Pearcy 1990; Kubiske, Abrams & Mostoller 1996). High irradiance had negative effects on Enchylaena carbon gain in very dry soils. Although most of the reduction in photosynthetic rate may be attributed to stomatal closure, observations of similar stomatal conductance between shaded and unshaded plants, yet different photosynthetic rates, suggests photoinhibitory effects or photodamage. For example, droughted Atriplex plants had higher gs in shade than in high light, yet Amax was higher in shaded conditions. High temperatures or water stress can have similar or interactive effects on the inactivation of the photosynthetic apparatus at high irradiance (Björkman & Powles 1984; Powles 1984). In this experiment, higher temperatures in unshaded conditions accompanied water stress so it is difficult to determine which factors were active. However, protection from excessive photon flux density under shaded conditions appears to be an important facilitation mechanism in arid systems and deserves further investigation.

When water was not limiting, Atriplex grown in high light had higher rates of photosynthesis than shade-grown plants. This was not found in the other two species. Similar light-limiting effects have been observed for other species commonly found in high-light environments (Kozlowski 1949; Holmgren 2000). Atriplex may thus be at a competitive disadvantage when water is abundant under trees, but not during drought. At the field site, Atriplex cover is significantly lower in the area adjacent to the trunk of dense-canopied Acacia trees than the outer canopy zone of sparse-canopied trees or immediately adjacent to trees (but not shaded) (Facelli & Brock 2000). Rhagodia and Enchylaena show the opposite pattern, having more cover in the most shaded microenvironment (dense inner canopy), and are very rarely found in habitats with higher light availability (open sites) (Facelli & Brock 2000). Although drought-tolerance characteristics do not appear to be disadvantageous to Atriplex in shade, stomatal sensitivity to light could restrict carbon gain and hence growth and competitive ability under shaded conditions, particularly in denser shade. There is no clear demarcation zone between shaded canopy microenvironments and the unshaded matrix microenvironment in arid systems, but rather a complex exists between the two (Holmgren et al. 1997; Moro et al. 1997). This is in addition to the variation found beneath trees associated with aspect (Hastwell 2001), canopy size (Tewksbury & Lloyd 2001) and canopy density (Facelli & Brock 2000). Gradients in light beneath different Acacia canopies will have complex effects on the distribution of these three shrubs.


On the basis of previously determined soil characteristics (Facelli & Brock 2000), it was predicted that after the onset of drought, soil moisture would remain available for longer in canopy soils. Therefore Ψs, gs and Amax would fall at slower rates in canopy soils than in open soils, but this generally did not occur. One explanation could be differences in shoot and root biomass between soil types that changed the demands on soil water. It was predicted that growth would be poorer in open soils because of its higher bulk density (N. Barnes, unpublished data). This would affect root penetration and water release (Greacen & Oh 1972; Hsiao 1973; Stirzaker, Passioura & Wilms 1996). Rhagodia and Enchylaena had higher biomass when grown in canopy soils and thus may have had higher water use when grown in these soils. In open soils, with lower root biomass, soils may have dried more slowly. Hence plants growing in canopy soils may not have had more soil water available for longer periods, as predicted.


Due to the controlling influence of water in arid systems, the effects of shade on the growth and water relations of plants are inextricably linked with the effects of water deficit. This study demonstrates that the effects of shade become positive with increasing water deficit. Extrapolation of responses of pot-grown seedlings to the field situation must necessarily be cautious, as pot soil temperatures are higher and pots dry out more quickly, therefore any regulatory mechanisms in response to water stress may not have time to be expressed (Ritchie 1981). However, the results suggest that, even over short time scales, acclimation to shade provided protection against water and/or heat stress. Atriplex showed the most varied response to shade in terms of biomass, Ψs, gs and Amax. This species appears to be tolerant of a broad range of conditions, reflected in its wide distributional range in field conditions. Rhagodia may be restricted to sites beneath canopies due to its high water usage. Stomatal conductance was very high in well watered plants growing in full sun and, as in Enchylaena, positive carbon gain was maintained for only a short time after the imposition of drought. Survivorship may therefore be reduced in open sites where water usage may be higher than in shaded sites because of higher evaporative demands. In contrast, Enchylaena may be restricted to sites beneath canopies due to the photoinhibitory effects of a combination of high photon flux densities, water stress and high temperatures in open sites during the summer months. Unlike Rhagodia, water usage by Enchylaena appears to be less affected by changes in light climate. Thus in this arid system these two species may be shade-dependent for different reasons.

This study has demonstrated how shade can positively affect the water relations and carbon gain of pot-grown seedlings during drought. Under field conditions, the positive benefits of plant shade must outweigh any negative effects of reduced soil moisture in these sites, compared to intercanopy spaces. Root competition from the overstorey species and interception of rainfall by the overstorey canopy may reduce soil moisture levels beneath canopies. However, soils beneath A. papyrocarpa canopies retain more water during warm, dry periods than intercanopy soils (Facelli & Brock 2000; Prider 2002), so the positive effects of canopies are twofold. During the warmer months they provide a moister microenvironment with a lower evaporative demand, and also protect plants from high radiation loads and high temperatures. The positive benefits of Acacia canopies may enable species such as Rhagodia and Enchylaena to persist in this habitat.


This is a contribution to the research program of the Middleback Field Centre. This research was funded by grants from the Australian Research Council and the Wildlife Conservation Fund (South Australian National Parks and Wildlife Council). We thank J. Watling for comments on the manuscript, R. Sinclair for the loan of equipment, and S. Gehrig and J. Goodfellow for their practical assistance.