Skip to main content
NCBI home page
Annals of Botany logoLink to Annals of Botany
. 2007 Oct 3;100(6):1297–1305. doi: 10.1093/aob/mcm226

Interactive Effects of Nutrient and Mechanical Stresses on Plant Morphology

Sara Puijalon 1,*, Jean-Paul Lena 1, Gudrun Bornette 1
PMCID: PMC2759263  PMID: 17913725

Abstract

Background and Aims

Plant species frequently encounter multiple stresses under natural conditions, and the way they cope with these stresses is a major determinant of their ecological breadth. The way mechanical (e.g. wind, current) and resource stresses act simultaneously on plant morphological traits has been poorly addressed, even if both stresses often interact. This paper aims to assess whether hydraulic stress affects plant morphology in the same way at different nutrient levels.

Methods

An examination was made of morphological variations of an aquatic plant species growing under four hydraulic stress (flow velocity) gradients located in four habitats distributed along a nutrient gradient. Morphological traits covering plant size, dry mass allocation, organ water content and foliage architecture were measured.

Key Results

Significant interactive effects of flow velocity and nutrient level were observed for all morphological traits. In particular, increased flow velocity resulted in size reductions under low nutrient conditions, suggesting an adaptive response to flow stress (escape strategy). On the other hand, moderate increases in flow velocity resulted in increased size under high nutrient conditions, possibly related to an inevitable growth response to a higher nutrient supply induced by water renewal at the plant surface. For some traits (e.g. dry mass allocation), a consistent sense of variation as a result of increasing flow velocity was observed, but the amount of variation was either reduced or amplified under nutrient-rich compared with nutrient-poor conditions, depending on the traits considered.

Conclusions

These results suggest that, for a given species, a stress factor may result, in contrasting patterns and hence strategies, depending on a second stress factor. Such results emphasize the relevance of studies on plant responses to multiple stresses for understanding the actual ecological breadth of species.

Key words: Allometry, Berula erecta, biomass partitioning, mechanical stress, morphology, multiple stresses, nutrient stress, phenotypic plasticity, submerged aquatic vegetation

INTRODUCTION

Under natural conditions, plants frequently encounter combinations of stress factors (Bazzaz, 1996; Sultan et al., 1998). Consequently, the individual ability to tolerate multiple stresses through morphological adjustments is a major feature that determines species survival and colonization, and hence the ecological breadth of the species (Chapin et al., 1987; Bazzaz, 1996; Sultan et al., 1998). Studies on plant responses to multiple stresses often deal only with stresses linked to limitations in several resources (nutrients, inorganic carbon or water availability, light quality or quantity; e.g. Urbas and Zobel, 2000; Sack, 2004), and focus on the optimization of resource acquisition and allocation.

In contrast, plant responses are largely unknown if one of the factors is not a resource stress, but a mechanical stress (e.g. wind or water motion). However, plants frequently experience resource and mechanical stresses simultaneously under natural conditions. In such conditions, plant growth is altered in a complex interactive manner (Idestam-Almquist and Kautsky, 1995; Power, 1996; Henry and Thomas, 2002) that cannot be simply predicted from the responses to each stress factor when considered independently.

In aquatic habitats, mechanical stress can result from exposure to water flow. Such stresses result ultimately in negative effects on plant growth (Idestam-Almquist and Kautsky, 1995; Power, 1996; Crossley et al., 2002). On the other hand, water motion can have an indirect positive effect on plant growth. Indeed, metabolic processes may be reduced in standing water by limited chemical flow across the boundary layer and under these conditions, moderate water motion can increase the flow of chemicals across the boundary layer (Hurd, 2000; Koch, 2001), thus favouring metabolic processes, and potentially plant growth (Madsen and Sondergaard, 1983; Koch, 1994; Thomas et al., 2000; Stewart and Carpenter, 2003). As chemical flux depends on nutrient concentrations in the mainstream water (Hurd, 2000; Thomas et al., 2000), the magnitude of this effect also depends on these concentrations.

Consequently, the aim of this study was to assess the combined effects of a mechanical stress (flow velocity) and a resource stress (nutrient availability) on the morphology of a higher aquatic plant species, Berula erecta, over a large range of both stresses.

In standing water, when exposed to increasing nutrient stress, plants tend to have a reduced total dry mass and leaf number (Zhang, 1996; Crossley et al., 2002), increased allocation to root and stem and decreased allocation to leaves (Gedroc et al., 1996; Madsen and Cedergreen, 2002), reduced water content of organs (Ryser, 1996; Craine et al., 2001), reduced leaflet number per leaf and reduced specific leaf area (SLA) (Li et al., 1999). These trait values correspond to adaptations usually observed in nutrient-poor habitats.

Morphological adaptations reducing damage risk have also been identified in plants exposed to mechanical stress. A reduced size and allocation to leaves exposed to flow, together with an increased allocation to below-ground organs (Idestam-Almquist and Kautsky, 1995; Bagger and Madsen, 2004) would result in weak forces and a greater anchorage efficiency, thus reducing the uprooting risk (Crook and Ennos, 1996; Niklas, 1998; Puijalon et al., 2005). Reduced leaf number and leaflet number per leaf (Puijalon and Bornette, 2006) could make the reconfiguration into a more streamlined shape easier (Pratt and Johnson, 2002). Moreover, reduced water content of organs and a lower SLA (Retuerto and Woodward, 1992; Boeger and Poulson, 2003), revealing thicker leaves and denser tissues could probably provide greater strength against breakage (Crook and Ennos, 1996; Niklas, 1999).

Ten morphological traits that were expected to be affected by both stress factors were consequently selected: total dry mass and leaf number, dry mass allocation to root, stem and leaves, water content of root, stem and leaves, number of leaflets per leaf and specific leaf area. The interactive effects of both nutrient and mechanical stresses were investigated. The hypothesis tested was that both stress factors do not affect plant traits independently but interact significantly for all plant traits, i.e. flow velocity should affect plant traits differently under low and high nutrient conditions.

This question was addressed by comparing the morphological traits of plants collected in natural habitats under 12 conditions corresponding to the combination of four nutrient levels and three flow velocity levels. This method was used because manipulating both factors using an experimental set-up (which is ideal for testing these hypotheses) was intractable in this case study as (a) nutrient levels were difficult to control accurately in running-water mesocosms and (b) transplants would be uprooted and fine-grained substrate scoured at the flow velocity required for the experiment.

MATERIALS AND METHODS

Species and study sites

Berula erecta (Hudson) Coville (Apiaceae, lesser water parsnip) is a perennial stoloniferous species, consisting of a rosette of petiolated dissected leaves. It measures 30–150 cm in height. The species shows a high morphological variability in size, leaflet number and leaf morphology (Puijalon and Bornette, 2004). It colonizes calcareous flowing habitats.

The study was carried out in four natural groundwater-supplied channels located in the Rhône River Basin (co-ordinates of the sites – 5°28′28″E, 45°98′87″N; 5°15′95″E, 45°80′50″N; 5°23′42″E, 45°83′72″N; 5°28′56″E, 45°98′63″N – for the four channels ranked according to increasing nutrient level). The channels were located in the same geological area and their bedload consisted of fluvio-glacial gravel and pebbles. They all included natural riffles with a permanent, relatively steady water flow. In each riffle of the four channels, plants were sampled in three patches of about 1 m2, ranging along a flow velocity gradient. Flow velocity was measured at the patch level, and the three patches were characterized by (a) a null velocity, i.e. <0·05 m s−1 (corresponding to the lower limit of flow measurement), (b) a low velocity, close to 0·15 m s−1 and (c) a high velocity, close to 0·25 m s−1). As the patches were close to each other (max 20 m), and as there was no source of nutrients in the sampled reaches, patches sampled in a given stream were considered similar in terms of water nutrient content. For this reason, the nutrient content of the water was assessed at the whole-riffle level.

Nutrient levels

The functioning of aquatic ecosystems is strongly governed by the nutrient content of water and substrate (Hansson et al., 1998; Khan and Ansari, 2005). In calcareous groundwater-supplied ecosystems, the composition of plant communities has been shown to depend mainly on the phosphate and ammonium concentrations in the water (Carbiener et al., 1990; Robach et al., 1996). The four selected channels ranged from oligotrophic (high nutrient stress) to meso-eutrophic (low nutrient stress). The selection was based on both (a) the occurrence of vegetation considered as representative of increasing trophic levels (plant communities in the four channels; see Supplementary Information Table available online) and (b) the phosphate and ammonium concentrations in the water (nutrient levels in the four channels; see Supplementary Information Fig. 1 available online). As aquatic plants can acquire nutrients by both roots and shoots (Madsen and Cedergreen, 2002), patches characterized by similar coarse substrate (gravels, pebbles) were selected in order to minimize a supplementary effect due to substrate characteristics (nutrient content or grain-size). Nutrient levels in the water were assessed through analyses of water samples collected in each channel on 14 dates spread over 4 months. Ammonium [N-NH3] and phosphate [PO43−] concentrations were measured by colorimetric means after standard HACH procedures (HACH Company, Loveland, CO, USA). As both nutrients rule plant communities, and as the authors did not want to discriminate between the quantitative roles of each component, the four channels were equally spaced between 1 (high nutrient stress, oligotrophic) and 4 (low nutrient stress, meso-eutrophic) along the nutrient availability gradient.

Flow velocity

Flow velocity was measured on each patch with a propeller (C2 current meter, OTT Messtechnik GmbH & Co. KG, Kempten, Germany) at a water depth 40 % above the substrate, which gives a good estimation of the average flow velocity in the water column (Dingman, 1984). Four random flow velocity measurements were made in each patch, avoiding hydraulic shelters (big pebbles, tall plants), on 15–17 dates depending on the channel, over a 5-month period, with contrasting river discharges.

When channels drain seepage water from the river, their discharges are related to the river discharge. In this case, the flow velocity effectively encountered by plants during the 4 months preceding plant sampling was extrapolated (no flood scouring occurred during this period). For this purpose, flow velocities measured on a given patch were regressed against daily river (Rhône or Ain River) discharge. This allowed the flow velocity really encountered by plants during the period preceding plant sampling to be evaluated precisely. However, in some cases (for the four standing patches and for two running patches under nutrient level 3), no significant relationship between flow velocity and river discharge was found. In these cases the flow velocity on the patch was estimated from the average value of the measurements done in situ. The average velocity encountered by plants was 0, 0·14 and 0·26 m s−1, respectively, in each velocity class (flow velocity encountered by plants during the 4-month period preceding sampling (see Supplementary Information Fig. 2 available online). The three velocity classes are thus referred to as V0, V0·14 and V0·26.

Other environmental characteristics

Water depth was 0·22 m on average and below 0·3 m for all patches except three (0·31, 0·41 and 0·70 m on average for these patches). Within each channel, flow-patches did not significantly differ for substrate grain size (gravels, pebbles) and other water characteristics. Indeed, patches were separated from each other by a few metres (from 1 m to 20 m), and channels drain groundwater that induced water renewal, even in standing patches, guaranteeing the homogeneity of physico-chemical characteristics among patches (temperature).

Plant sampling

Berula erecta is evergreen and totally submerged during winter, whereas it becomes emerged in standing water during summer. Plant sampling was therefore carried out in early spring, from 11 February to 10 April 2002. At that time, plants would have grown totally submerged and fully encountered the nutrient and velocity conditions of the habitats for several months. The extended duration of sampling was due to the large number of plant traits measured. However, in order to avoid as far as possible any confounding effect of ramet growth during sampling, the different patches were sampled randomly during the two months that were necessary for plant trait measurements. Thirty fully submerged individuals (ramets) were collected from distinct clones in each flow patch. A ramet of B. erecta was defined as a single rooted rosette. Any horizontal stolon growing from the ramet was removed. After collection, the plants were stored in aerated tap water at 16 °C for a maximum of 2 d until measurements were taken.

Morphometry

The following traits were measured on each plant: (a) number of leaves and number of leaflets per leaf; (b) plant mass (g) – plants were divided into roots, stems, petioles and leaflets and the different parts were weighed to obtain fresh and dry mass (measured after drying for 48 h at 85 °C); (c) leaf area (cm2) – leaves were scanned (150 dpi, Epson Expression 1680 scanner) and the images were analysed with WinFolia 2001 image analysis software (Regent Instrument Inc., Quebec, Canada). Leaf area was measured for the whole plant.

These measurements were used to calculate a data set consisting of ten morphological traits that were divided into four groups.

  1. Plant size was assessed by total dry mass and leaf number. As plant height, total leaf area and total dry mass were highly correlated (r = 0·78, P < 10−4 and r = 0·89, P < 10−4 for the correlation between total dry mass and plant height and total leaf area, respectively), overall plant size was described only by its total dry mass.

  2. Dry mass allocation: allocations to different plant parts were described by dry mass allocations to roots, stem and leaves relative to total plant mass.

  3. Organ water content: for each organ (root, stem and leaf), water content was assessed by expressing fresh mass relative to dry mass.

  4. Traits related to foliage architecture: foliage architecture was described by specific leaf area (SLA = leaf area/leaf dry mass) and number of leaflets relative to leaf number.

Statistical analysis

All mass and surface variables were loge transformed prior to analysis to improve the normality of residuals and homogeneity of variance.

First, each trait was tested to find out if the interactive effect of nutrient level and flow velocity was significant. For plant dry mass, the effects of both stress factors and their interaction were assessed through a two-way ANOVA, treating both stress factors and their interaction as nominal explicative terms (GLM procedure; SAS Institute, 2004).

The number of leaves per ramet was a discrete ordinal variable. Consequently, its variation pattern was assessed with a multinomial model using a cumulative logit (GENMOD procedure; SAS Institute, 2004) (Agresti, 2002). Plants with two leaves (four individuals) and three leaves (31 individuals) were grouped in the same class (≤3 leaves) and plants with seven (five individuals), eight (two individuals) and nine leaves (one individual) were grouped in the ≥7 leaves class, in order to homogenize individual distributions within classes.

For all other traits, the effects of both stress factors and their interaction were studied using allometric relationships, which allow comparisons at a common plant size (Coleman et al., 1994; Lenssen et al., 2003). ANCOVA was used to test how environmental factors affect relationships between traits, rather than comparing the ratio of traits, which often depend on plant size (Coleman et al., 1994). An ANCOVA (GLM procedure; SAS Institute, 2004) was carried out on the dry mass allocation of roots, stem and leaves relative to total dry mass, water content of the three compartments and SLA. For dry mass allocation, the ANCOVA was performed for each compartment, using the dry mass of the compartment (root, stem and leaves) as a dependent variable and total dry mass as a covariate. To compare the water content, for each compartment, the ANCOVA was performed on the fresh mass as dependent variable and the dry mass as covariate. For SLA, the ANCOVA was performed on the leaf area treated as the dependent variable and the leaf dry mass as the covariate. The number of leaflets per leaf was studied with a Poisson model, adapted for counting data (McCullagh and Nelder, 1989). In this model, leaflet number was treated as the dependent variable and leaf number as covariate.

For all analyses, interaction terms between the two stress factors and the covariate (if present) were first introduced in the model. Non-significant higher-order interaction terms were then stepwise removed to obtain the final model (type III sums of squares; SAS Institute, 2004). If a significant interaction between a stress factor and the covariate occurred, then the response was examined at the first, second, and third quartiles of the covariate to describe the interaction.

To investigate the nature of interaction, first the effect of increasing nutrient stress was assessed. For this purpose, contrast analyses were used to test orthogonal polynomial terms describing the linear, quadratic and cubic components in the trend. For the sake of clarity, only the linear component in the trend is presented.

Then the effect of increasing flow velocity was investigated to determine whether the flow velocity effect on plant traits was reduced under high nutrient conditions. Within a given nutrient level, differences between trait values for different flow velocities were tested using contrast tests on least squares means (i.e. predicted values at the levels of nominal effects). To study the variation trend of a trait for increasing velocity, contrast analyses were used to test orthogonal polynomial terms describing the linear and quadratic components in the trend. Results of these tests are indicated only when a pattern of variation could be identified.

RESULTS

Total dry mass, leaf number, root allocation and leaf water content were significantly affected by both stress factors and their interaction (Table 1). Stem allocation and stem water content were significantly affected by both stress factors and their interaction, as well as the interaction between the covariate and nutrient level (Table 1). Leaf allocation, root water content, SLA and leaflet number per leaf were all significantly affected by the three-way interaction between both stress factors and the covariate (Table 1). For leaflet number per leaf, the second quartile was not an integer (4·5 leaves) and was therefore not considered for the following analyses.

Table 1.

Relationships between plant morphological traits and flow velocity, nutrient level and their interactions

Plant size
 Dry mass allocation
Total dry mass
 Leaf number
 Root mass ratio
 Stem mass ratio
 Leaf mass ratio
Dependent variable Log(total dry mass) Leaf number Log(root dry mass) Log(stem dry mass) Log(leaf dry mass)
Covariate Log(total dry mass) Log(total dry mass) Log(total dry mass)
d.f. F P χ2 P F P F P F P
Nutrient level (N) 3 261·1 <0·0001 45·1 <0·0001 104·3 <0·0001 6·6 0·0002 13·4 <0·0001
Flow velocity (V) 2 88·9 <0·0001 42·3 <0·0001 39·2 <0·0001 26 <0·0001 3·1 0·047
Covariate (C) 1 459·2 <0·0001 3550·9 <0·0001 2947·2 <0·0001
N×V 6 31·9 <0·0001 22 0·001 7·3 <0·0001 3·8 0·001 5·1 <0·0001
N×C 3 1·7 0·17 3·7 0·01 2·8 0·04
V×C 2 1·1 0·33 1·5 0·22 3·6 0·03
N×V×C 6 0·3 0·92 1 0·43 3·4 0·003
Organ water content Foliage architecture
Root water content Stem water content Leaf water content SLA Number of leaflets per leaf
Dependent variable Log(root fresh mass) Log(stem fresh mass) Log(leaf fresh mass) Log(leaf area) Leaflet number
Covariate Log(root dry mass) Log(stem dry mass) Log(leaf dry mass) Log(leaf dry mass) Leaf number
d.f. F P F P F P F P χ2 P
Nutrient level (N) 3 4·7 0·003 6·9 0·001 69·5 <0·0001 13·4 <0·0001 17 0·0007
Flow velocity (V) 2 0·9 0·39 5·2 0·002 51·2 <0·0001 3·1 0·047 0·6 0·72
Covariate (C) 1 1279·1 <0·0001 1709·1 <0·0001 3124·8 <0·0001 2947·2 <0·0001 235·9 <0·0001
N×V 6 3·5 0·002 6·9 <0·0001 17·7 <0·0001 5·1 <0·0001 40 <0·0001
N×C 3 9·6 <0·0001 3·6 0·01 2·4 0·07 2·8 0·04 13·9 0·003
V×C 2 1·3 0·29 0·3 0·72 0·5 0·62 3·6 0·03 2·3 0·32
N×V×C 6 4·9 <0·0001 1·9 0·08 1·86 0·09 3·4 0·003 15·8 0·015
Open in a new tab

Tests carried out were two-way ANOVA for total dry mass and leaf number, and two-way ANCOVA for dry mass allocation and water content of the three compartments, SLA and number of leaflets per leaf. Non-significant interaction terms that were successively dropped are indicated in italics (F and P-value correspond to the model with higher-order interaction where the term was present).

Effect of nutrient stress in standing water

For velocity V0, total dry mass, leaf number, leaf allocation, stem and leaf water content and SLA increased, whereas root and stem allocation decreased, with increasing nutrient level (Figs 14 and Table 2). Two traits (root water content and leaflet number per leaf) varied in a complex way with nutrient level. Even though root water content tended to increase with increasing nutrient level (Table 2), it varied in a non-monotonic way for increasing nutrient level due to the large significance of the quadratic and cubic terms in the trend (Fig. 3A, e.g. for the second quartile of root fresh mass, F1,359 = 14·2, P < 0·001 and F1,359 = 45·4, P < 0·001 for quadratic and cubic term, respectively).

Fig. 1.

Fig. 1.

Open in a new tab

Size traits according to nutrient level and flow velocity; (A) total dry mass and (B) leaf number. Points indicate least squares means [loge(total dry mass)] ± s.e. for total dry mass and log(odd) ± s.e. at a mean cut point for number of leaves. Odds for an event are the ratio of the probability of the event occurring and the probability of that event not occurring and represent the chance for a plant to have a higher leaf number.

Fig. 2.

Fig. 2.

Open in a new tab

Dry mass allocations (A, root; B, stem; C, leaf dry mass relative to total dry mass) according to nutrient level and flow velocity, shown as least-squares means [loge(organ dry mass)] ± s.e., with ‘organ’ being roots, stems and leaves, respectively. Points are represented for each quartile of the total dry mass, when at least one interaction between stress factor and covariate was significant.

Fig. 3.

Fig. 3.

Open in a new tab

Water content of the three plant compartments (A, root; B, stem; C, leaf water content) according to nutrient level and flow velocity, shown as least squares means [loge(organ dry mass)] ± s.e., with ‘organ’ being roots, stems and leaves, respectively. Points are represented for each quartile of the total dry mass, when at least one interaction between stress factor and covariate was significant.

Fig. 4.

Fig. 4.

Open in a new tab

Traits describing foliage architecture according to nutrient level and flow velocity: (A) SLA, shown as least squares means loge(leaf area) ± s.e. and (B) number of leaflets per leaf ± s.e.. Points are represented for each quartile of the total dry mass, when at least one interaction between stress factor and covariate was significant.

Table 2.

Contrast analyses testing a linear term for traits as nutrient level increases, carried out at velocity V0

Trait Quartile of covariate F1,359 or χ12 Sense of variation
Total dry mass 60·9*** +
Leaf number 18·6*** +
Root allocation 302·5*** −
Stem allocation 1st q. (total dry mass) 26·8*** −
2nd q. (total dry mass) 53·2*** −
3rd q. (total dry mass) 44·7*** −
Leaf allocation 1st q. (total dry mass) 60·1*** +
2nd q. (total dry mass) 111·5*** +
3rd q. (total dry mass) 86·9*** +
Root water content 1st q. (root fresh mass) 9·4** +
2nd q. (root fresh mass) 15·2*** +
3rd q. (root fresh mass) 10·1** +
Stem water content 1st q. (stem fresh mass) 160·2*** +
2nd q. (stem fresh mass) 234·8*** +
3rd q. (stem fresh mass) 228·0*** +
Leaf water content 152·3*** +
SLA 1st q. (leaf dry mass) 33·0*** +
2nd q. (leaf dry mass) 198·1*** +
3rd q. (leaf dry mass) 181·0*** +
Leaflet number per leaf 1st q. (leaf number) 0·3 n.s.
3rd q. (leaf number) 2·46 n.s.
Open in a new tab

Results indicate F1,359 or χ12 (for number of leaves and number of leaflets per leaf), the significance level and the sense of variation of the trait according to nutrient level. Tests were carried out for each quartile (1st, 2nd and 3rd q.) of the covariate (if present), when at least one interaction between stress factor and covariate was significant. ***P <0·001; **P < 0·01; *P < 0·05; n.s., not significant.

Contrast analyses testing the quadratic and cubic terms have also been carried out but are not shown, as they do not correspond to hypotheses.

Effect of flow velocity

Flow velocity effects were identified for six morphological traits (total dry mass, leaf number, root and leaf allocation, leaf water content and leaflet number per leaf). For all these traits, the flow velocity effect differed between nutrient levels.

Total dry mass

At the third nutrient level, total dry mass was significantly higher at velocities V0·14 and V0·26 than at velocity V0 (Fig. 1A, contrast test, velocity V0 vs. velocities V0·14 and V0·26: F1,359 = 7·0, P = 0·009). For the other nutrient levels, total dry mass was significantly lower for velocities V0·14 and V0·26 than for velocity V0 (Fig. 1A, contrast test, velocity V0 vs. velocities V0·14 and V0·26 for nutrient levels 1, 2 and 4, respectively: F1,359 = 108·4, P < 0·0001; F1,359 = 127·9, P < 0·0001; F1,359 = 12·8, P = 0·0004).

Leaf number

Leaf number did not differ significantly between flow velocities for the third nutrient level (contrast test, velocity V0 vs. velocities V0·14 and V0·26: χ12 = 2·5, n.s.), whereas it was significantly lower for velocities V0·14 and V0·26 than for velocity V0 for all other nutrient levels (Fig. 1B, contrast test, velocity V0 vs. velocities V0·14 and V0·26 for nutrient levels 1, 2 and 4, respectively: χ12 = 11·0, P < 0·0001; χ12 = 4·2, P = 0·04; χ12 = 36·3, P <0·0001). For velocity V0·26, leaf number did not differ significantly between nutrient levels (Fig. 1B, χ32 = 1·5, P = 0·7).

Root allocation

Root allocation did not differ between flow velocities for the first nutrient level (contrast test, velocity V0 vs. velocities V0·14 and V0·26: F1,359 = 0·16, n.s.), but for other nutrient levels, it was significantly lower for velocities V0·14 and V0·26 than for velocity V0 (Fig. 2A, contrast test, velocity V0 vs. velocities V0·14 and V0·26 for nutrient levels 2, 3 and 4, respectively: F1,359 = 30·5, P <0·0001; F1,359 = 30·8, P < 0·0001; F1,359 = 64·4, P < 0·0001). This effect of flow velocity increased when nutrient levels rose (Fig. 2A, contrast test, difference between V0 and velocities V0·14 and V0·26 according to the nutrient level: F1,359 = 21·3, P < 0·0001; F1,359 = 2·9, n.s., for linear and quadratic terms, respectively).

Leaf allocation

Leaf allocation was significantly lower for velocities V0·14 and V0·26 than for velocity V0 for all but the fourth nutrient level (Fig. 2C, contrast test, velocity V0 vs. velocities V0·14 and V0·26 for nutrient levels 1, 2 and 3, respectively: F1,359 = 17·6, P < 0·0001; F1,359 = 20·2, P < 0·0001; F1,359 = 11·6, P = 0·0008). For the fourth nutrient level, leaf allocation was independent of flow velocity (contrast test, velocity V0 vs. velocities V0·14 and V0·26: F1,359 = 3·8, n.s.). The difference in leaf allocation decreased as nutrient levels rose (Fig. 2C, contrast test, difference between V0 vs. velocities V0·14 and V0·26 according to the nutrient level: F1,359 = 7·9, P = 0·005; F1,359 = 0·5, n.s., for linear and quadratic terms, respectively).

Leaf water content

Leaf water content was significantly higher at velocity V0 than at velocities V0·14 and V0·26, for all but the second nutrient level (Fig. 3C, contrast test, velocity V0 vs. velocities V0·14 and V0·26 for nutrient levels 1, 2, 3 and 4, respectively: F1,359 = 102·3, P < 0·0001; F1,359 = 1·5, n.s.; F1,359 = 30·0, P < 0·0001; F1,359 = 50·7, P < 0·0001).

Number of leaflets per leaf

For the lowest two nutrient levels, the number of leaflets per leaf was significantly higher at velocity V0 compared with velocities V0·14 and V0·26, whatever the leaf number (Fig. 4B contrast test, velocity V0 vs. velocities V0·14 and V0·26 for nutrient levels 1 and 2, respectively: χ12 = 45·3, P < 0·0001 and χ12 = 51·3, P < 0·0001 for low leaf number and χ12 = 44·2, P < 0·0001 and χ12 = 32·3, P < 0·0001 for high leaf number). However, the opposite trend was observed for the highest two nutrient levels (Fig. 4B, contrast test, velocity V0 vs. velocities V0·14 and V0·26 for nutrient levels 3 and 4, respectively: χ12 = 58·9, P < 0·0001 and χ12 = 31·3, P < 0·0001 for low leaf number and χ12 = 191·2, P < 0·0001 and χ12 = 40·9, P < 0·0001 for high leaf number).

No particular flow velocity effect could be identified on the other four traits (stem allocation, root and stem water content, and SLA; Figs 24), as the flow velocity effect on those traits varied in a complex way according to nutrient level.

DISCUSSION

In accordance with our hypothesis, all traits were significantly affected by both stress factors and their interaction, leading to a complex effect of both environmental factors for some traits. Contrary to the present results, previous studies only demonstrated the interactive effects of resource and mechanical stresses on plant size, but not (or only weakly significant) for other morphological traits (Idestam-Almquist and Kautsky, 1995; Crossley et al., 2002; Henry and Thomas, 2002). This could result from the contrasting ranges of values used for the abiotic (mechanical and resource) factors in the different studies. In the present study, both nutrient levels and flow velocity exhibited rather large variations. This experimental design (at least three levels of stress) allowed the delineation of the complex, non-linear interaction between hydraulic and nutrient stresses that cannot otherwise be properly addressed without.

Effect of nutrient stress

In the present study, trait variations observed along the nutrient gradient were consistent with previous studies, even though root water content and leaflet number per leaf were not related in a simple way to nutrient level. Lower total dry mass, leaf allocation and leaf number, higher allocation to below-ground organs and reduced SLA have been frequently observed for individuals growing in low-nutrient habitats either for terrestrial (Gedroc et al., 1996; Li et al., 1999; McConnaughay and Coleman, 1999; Crossley et al., 2002) or aquatic plants (Idestam-Almquist and Kautsky, 1995; Crossley et al., 2002; Madsen and Cedergreen, 2002; Xie et al., 2005). Moreover, a lower water content (denser tissues) has been observed among species colonizing nutrient-poor habitats (Ryser, 1996; Craine et al., 2001).

Effect of flow velocity

In accordance with our hypothesis, the effect of flow velocity differed markedly according to nutrient level for all morphological traits. However, the nature of the interactive effect of flow velocity and nutrient level differed between traits. For six traits, it was possible to interpret the pattern of variations in response to flow velocity, whereas for four traits (stem allocation, SLA and root and stem water content), this pattern was too complex to be interpreted.

Effects of flow velocity at low nutrient level

For five traits (total dry mass, leaf number, leaf allocation, leaflet number per leaf and leaf water content), at low nutrient levels the expected variation pattern was observed, based on adaptation to mechanical stress (i.e. reduction of total dry mass, leaf number, leaf allocation, leaflet number per leaf and leaf water conten; Retuerto and Woodward, 1992; Boeger and Poulson, 2003; Puijalon and Bornette, 2004), possibly because the direct effect of mechanical stress on plant morphology was predominant. When nutrient levels in the mainstream water are low, the nutrient uptake is limited due to the low nutrient concentration, whatever the level of water stirring. Under such conditions, the beneficial effect of water stirring on plant growth is negligible.

In contrast to previous studies (Retuerto and Woodward, 1992; Idestam-Almquist and Kautsky, 1995; Niklas, 1998), mechanical stress did not result in an increased allocation to roots. For this trait, nutrient stress had such a strong effect that the occurrence of a new stress had no additional effect. Such reduced effects had already been reported for other kinds of stress (Chapin et al., 1987; Lenssen et al., 2003).

Effects of flow velocity at high nutrient level

For four traits (total dry mass, leaf number, leaf allocation and leaflet number per leaf), both stress factors interacted in such a way that the effect of flow velocity was reduced and even reversed for dry mass and leaflet number per leaf under high nutrient conditions compared with nutrient-poor conditions.

However, the flow velocity effect also differed between the 3rd and the 4th nutrient levels. At the 3rd nutrient level, the flow velocity effect was reduced for leaf number and leaf allocation, and reversed for total dry mass and number of leaflets per leaf compared with nutrient-poor conditions, suggesting that indirect beneficial effect of water motion on plant growth probably exceeded the direct effect of mechanical stress (Crossley et al., 2002). Moderate water stirring probably reduced the well-developed concentration gradient between the uptaking plant surface and the mainstream water and consequently led to enhanced nutrient and carbon availability. With very high nutrient levels (4th nutrient level), the beneficial indirect water motion effect was less pronounced than at nutrient level 3 (and even negligible for leaf number). At this nutrient level, the indirect water motion effect is probably negligible because the plant cannot take up the nutrients fast enough to create a nutrient gradient. On the other hand, for a second group of traits (i.e. root allocation and leaf water content), a high nutrient level did not lead to a reduced flow stress effect.

Consequences for plant function in stressful environments

These results raise the issue of the adaptive value of responses to hydraulic stress. At the highest flow velocity, leaf numbers converged to a similar value, whatever the nutrient level. This value could represent, for that species, a morphological optimal for spatial organization of the above-ground biomass, improving the plant's capacity to compact into a more streamlined shape through reconfiguration (Pratt and Johnson, 2002).

The present results demonstrate also that, at the organism level, the interactive effect of both factors can be strong enough to result in contrasting and even opposed responses to water movement under low and high nutrient conditions. Some trait variations (e.g. plant size reduction, or leaf allocation) are beneficial to escape flow stress in nutrient-poor situations for that species (Puijalon et al., 2005). However, the pattern observed in nutrient-rich situations could be a case of inevitable response (see Sultan, 2000), elicited by enhanced chemical supply. Such trait variations (particularly larger size) either increase the risk of uprooting and damage, or plants can show other morphological adjustments that reduce plant drag and allow rapid recovery after damage. Some of the adaptations that possibly occur are enhanced leaf turn-over or the expression of morphological traits that reduce drag without significantly reducing plant size.

Conclusions

Water motion is a major factor determining plant growth and morphology. This study demonstrated, on established populations, the strong interactive effect of mechanical stress and nutrient level on plant traits. Moreover, several patterns of interaction were identified between both stress factors, depending on the trait considered, which impede the establishment of a general model. The next step would be to determine the origin of the patterns observed; either they resulted, at least in part, from plastic responses, or different reactions norms could have been selected.

SUPPLEMENTARY INFORMATION

The following Supplementary Information is available online at http://aob.oxfordjournals.org/. Table: plant communities in the four channels; species abundance results from the average values of Braun–Blanquet abundance cover indices are given for the whole samples. Fig. 1: Nutrient levels in the four channels. Fig. 2: flow velocity encountered by plants during the 4-month period preceding sampling are given.

ACKNOWLEDGEMENTS

We thank D. Reynaud and E. Malet for their technical assistance. This study was partly funded by the ‘ACI Ecologie Quantitative’ of the MENRT (Ministère de l'Education Nationale, de la Recherche et de la Technologie), and was carried out under the aegis of the long-term ecological research programme on the Rhône River Basin (Zone Atelier Bassin du Rhône). We thank Professor Hideyuki Takahashi, the anonymous reviewers and the journal editor for fruitful comments on a previous version of the manuscript.

LITERATURE CITED

  1. Agresti A. Categorical data analysis. New York, NY: John Wiley & Sons; 2002. [Google Scholar]
  2. Amoros C, Bornette G, Henry CP. A vegetation-based method for ecological diagnosis of riverine wetlands. Environmental Management. 2000;25:211–227. doi: 10.1007/s002679910017. [DOI] [PubMed] [Google Scholar]
  3. Bagger J, Madsen TV. Morphological acclimation of aquatic Littorella uniflora to sediment CO2 concentration and wave exposure. Functional Ecology. 2004;18:946–951. [Google Scholar]
  4. Bazzaz FA. Plants in changing environments: linking physiological, population and community ecology. Cambridge: Cambridge University Press; 1996. [Google Scholar]
  5. Boeger MRT, Poulson ME. Morphological adaptations and photosynthetic rates of amphibious Veronica anagallis-aquatica L. (Scrophulariaceae) under different flow regimes. Aquatic Botany. 2003;75:123–135. [Google Scholar]
  6. Carbiener R, Trémolières M, Mercier JL, Ortscheit A. Aquatic macrophyte communities as bioindicators of eutrophication in calcareous oligosaprobe stream waters (Upper Rhine plain, Alsace) Vegetatio. 1990;86:71–88. [Google Scholar]
  7. Chapin FS, III, Bloom AJ, Field CB, Waring RH. Plant responses to multiple environmental factors. Bioscience. 1987;37:49–57. [Google Scholar]
  8. Coleman JS, McConnaughay KDM, Ackerly DD. Interpreting phenotypic variation in plants. Trends in Ecology and Evolution. 1994;9:187–191. doi: 10.1016/0169-5347(94)90087-6. [DOI] [PubMed] [Google Scholar]
  9. Craine JM, Froehle J, Tilman DG, Wedin DA, Chapin FS., III The relationships among root and leaf traits of 76 grassland species and relative abundance along fertility and disturbance gradients. Oikos. 2001;93:274–285. [Google Scholar]
  10. Crook MJ, Ennos AR. Mechanical differences between free-standing and supported wheat plants Triticum aestivum L. Annals of Botany. 1996;77:197–202. [Google Scholar]
  11. Crossley MN, Dennison WC, Williams RR, Wearing AH. The interaction of water flow and nutrients on aquatic plant growth. Hydrobiologia. 2002;489:63–70. [Google Scholar]
  12. Dingman SL. Fluvial hydrology. New York, NY: W. H. Freeman and Co; 1984. [Google Scholar]
  13. Gedroc JJ, McConnaughay KDM, Coleman JS. Plasticity in root/shoot partitioning: optimal, ontogenetic, or both? Functional Ecology. 1996;10:44–50. [Google Scholar]
  14. Hansson L-A, Annadotter H, Bergman E, Hamrin SF, Jeppesen E, Kairesalo T, et al. Biomanipulation as an application of food-chain theory: constraints, synthesis, and recommendations. Ecosystems. 1998;1:558–574. [Google Scholar]
  15. Henry HAL, Thomas SC. Interactive effects of lateral shade and wind on stem, allometry, biomass allocation, and mechanical stability in Abutilon theophrasti (Malvaceae) American Journal of Botany. 2002;89:1609–1615. doi: 10.3732/ajb.89.10.1609. [DOI] [PubMed] [Google Scholar]
  16. Hurd CL. Water motion, marine macroalgal physiology, and production. Journal of Phycology. 2000;36:453–472. doi: 10.1046/j.1529-8817.2000.99139.x. [DOI] [PubMed] [Google Scholar]
  17. Idestam-Almquist J, Kautsky L. Plastic responses in morphology of Potamogeton pectinatus L. to sediment and above-sediment conditions at two sites in the northern Baltic proper. Aquatic Botany. 1995;52:205–216. [Google Scholar]
  18. Khan FA, Ansari AA. Eutrophication: an ecological vision. Botanical Review. 2005;71:449–482. [Google Scholar]
  19. Koch EW. Hydrodynamics, diffusion-boundary layers and photosynthesis of the seagrasses Thalassia testudinum and Cymodocea nodosa. Marine Biology. 1994;118:767–776. [Google Scholar]
  20. Koch EW. Beyond light: physical, geological, and geochemical parameters as possible submersed aquatic vegetation habitat requirements. Estuaries. 2001;24:1–17. [Google Scholar]
  21. Lenssen JPM, Menting FBJ, Van Der Putten WH. Plant response to simultaneous stress of waterlogging and shade: amplified or hierarchical effects? New Phytologist. 2003;157:281–290. doi: 10.1046/j.1469-8137.2003.00666.x. [DOI] [PubMed] [Google Scholar]
  22. Li B, Suzuki J-I, Hara T. Competitive ability of two Brassica varieties in relation to biomass allocation and morphological plasticity under varying nutrient availability. Ecological Research. 1999;14:255–266. [Google Scholar]
  23. McConnaughay KDM, Coleman JS. Biomass allocation in plants: ontogeny or optimality? A test along three resource gradients. Ecology. 1999;80:2581–2593. [Google Scholar]
  24. McCullagh P, Nelder JA. Generalized linear models. London: Chapman and Hall; 1989. [Google Scholar]
  25. Madsen TV, Cedergreen N. Sources of nutrients to rooted submerged macrophytes growing in a nutrient-rich stream. Freshwater Biology. 2002;47:283–291. [Google Scholar]
  26. Madsen TV, Sondergaard M. The effects of current velocity on the photosynthesis of Callitriche stagnalis Scop. Aquatic Botany. 1983;15:187–193. [Google Scholar]
  27. Niklas KJ. Effects of vibration on mechanical properties and biomass allocation pattern of Capsella bursa-pastoris (Cruciferae) Annals of Botany. 1998;82:147–156. [Google Scholar]
  28. Niklas KJ. Research review: a mechanical perspective on foliage leaf form and function. New Phytologist. 1999;143:19–31. [Google Scholar]
  29. Power P. Effects of current velocity and substrate composition on growth of Texas wildrice (Zizania texana) Aquatic Botany. 1996;55:199–204. [Google Scholar]
  30. Pratt MC, Johnson AS. Strength, drag, and dislodgment of two competing intertidal algae from two wave exposures and four seasons. Journal of Experimental Marine Biology and Ecology. 2002;272:71–101. [Google Scholar]
  31. Puijalon S, Bornette G. Morphological variation of two taxonomically distant plant species along a natural flow velocity gradient. New Phytologist. 2004;163:651–660. doi: 10.1111/j.1469-8137.2004.01135.x. [DOI] [PubMed] [Google Scholar]
  32. Puijalon S, Bornette G. Phenotypic plasticity and mechanical stress: biomass partitioning and clonal growth of an aquatic plant species. American Journal of Botany. 2006;93:1090–1099. doi: 10.3732/ajb.93.8.1090. [DOI] [PubMed] [Google Scholar]
  33. Puijalon S, Bornette G, Sagnes P. Adaptations to increasing hydraulic stress: morphology, hydrodynamics and fitness of two higher aquatic plant species. Journal of Experimental Botany. 2005;56:777–786. doi: 10.1093/jxb/eri063. [DOI] [PubMed] [Google Scholar]
  34. Retuerto R, Woodward FI. Effects of windspeed on the growth and biomass allocation of white mustard Sinapis alba L. Oecologia. 1992;92:113–123. doi: 10.1007/BF00317271. [DOI] [PubMed] [Google Scholar]
  35. Robach F, Thiébaut G, Trémolières M, Muller S. A reference system for continental running waters: plant communities as bioindicators of increasing eutrophication in alkaline and acidic waters in north-east France. Hydrobiologia. 1996;340:67–76. [Google Scholar]
  36. Ryser P. The importance of tissue density for growth and life span of leaves and roots: a comparison of five ecologically contrasting grasses. Functional Ecology. 1996;10:717–723. [Google Scholar]
  37. Sack L. Responses of temperate woody seedlings to shade and drought: do trade-offs limit potential niche differentiation? Oikos. 2004;107:110–127. [Google Scholar]
  38. SAS Institute. Cary, NC, USA: SAS Institute Inc; 2004. SAS OnlineDoc®, Version 9. [Google Scholar]
  39. Stewart HL, Carpenter RC. The effects of morphology and water flow on photosynthesis of marine macroalgae. Ecology. 2003;84:2999–3012. [Google Scholar]
  40. Sultan SE. Phenotypic plasticity for plant development, function and life-history. Trends in Plant Science. 2000;5:537–542. doi: 10.1016/s1360-1385(00)01797-0. [DOI] [PubMed] [Google Scholar]
  41. Sultan SE, Wilczek AM, Bell DL, Hand G. Physiological response to complex environments in annual Polygonum species of contrasting ecological breadth. Oecologia. 1998;115:564–578. doi: 10.1007/s004420050554. [DOI] [PubMed] [Google Scholar]
  42. Thomas FIM, Cornelisen CD, Zande JM. Effects of water velocity and canopy morphology on ammonium uptake by seagrass communities. Ecology. 2000;81:2704–2713. [Google Scholar]
  43. Urbas P, Zobel K. Adaptive and inevitable morphological plasticity of three herbaceous species in a multi-species community: field experiment with manipulated nutrients and light. Acta Oecologica. 2000;21:139–147. [Google Scholar]
  44. Xie Y, An S, Yao X, Xiao K, Zhang C. Short-time response in root morphology of Vallisneria natans to sediment type and water-column nutrient. Aquatic Botany. 2005;81:85–96. [Google Scholar]
  45. Zhang J. Interactive effects of soil nutrients, moisture and sand burial on the development, physiology, biomass and fitness of Cakike dentula. Annals of Botany. 1996;78:591–598. [Google Scholar]

Articles from Annals of Botany are provided here courtesy of Oxford University Press

RESOURCES