|Home | About | Journals | Submit | Contact Us | Français|
The plastic alterations of clonal architecture are likely to have functional consequences, as they affect the spatial distribution of ramets over patchy environments. However, little is known about the effect of mechanical stresses on the clonal growth. The aim of the present study was to investigate the clonal plasticity induced by mechanical stress consisting of continuous water current encountered by aquatic plants. More particularly, the aim was to test the capacity of the plants to escape this stress through clonal plastic responses.
The transplantation of ramets of the same clone in two contrasting flow velocity conditions was carried out for two species (Potamogeton coloratus and Mentha aquatica) which have contrasting clonal growth forms. Relative allocation to clonal growth, to creeping stems in the clonal biomass, number and total length of creeping stems, spacer length and main creeping stem direction were measured.
For P. coloratus, plants exposed to water current displayed increased total length of creeping stems, increased relative allocation to creeping stems within the clonal dry mass and increased spacer length. For M. aquatica, plants exposed to current displayed increased number and total length of creeping stems. Exposure to current induced for both species a significant increase of the proportion of creeping stems in the downstream direction to the detriment of creeping stems perpendicular to flow.
This study demonstrates that mechanical stress from current flow induced plastic variation in clonal traits for both species. The responses of P. coloratus could lead to an escape strategy, with low benefits with respect to sheltering and anchorage. The responses of M. aquatica that may result in a denser canopy and enhancement of anchorage efficiency could lead to a resistance strategy.
A high proportion of all plant species are clonal, consisting of networks of interconnected ramets (Klimes et al., 1997). Many among them have been demonstrated to alter their morphology in response to environmental factors, both at the level of individual ramets and at the level of the clone (i.e. set of interconnected ramets). Such plastic responses may include both alterations of allocations between the different parts of the clone (e.g. spacer versus ramets) and modifications of clonal architecture (e.g. number and length of stolons/rhizomes, spacer lengths defined as the distance between consecutive ramets, branching patterns, sequence of the ramets; de Kroon and Hutchings, 1995; Valverde and Pisanty, 1999; Ye et al., 2006; Ikegami et al., 2007). These alterations of clonal growth are likely to have functional consequences and to affect the spatial distribution of ramets over patchy environments. Particularly, dense growth forms (phalanx growth forms; Lovett-Doust, 1981) due to reduced spacer length and increased branching could allow plants to occupy favourable patches in a heterogeneous environment (de Kroon et al., 1994; Dong and de Kroon, 1994; de Kroon and Hutchings, 1995). In contrast, dispersed growth forms (guerrilla growth form; Lovett-Doust, 1981) due to increased spacer length and reduced branching could allow the plant to escape from less-favourable conditions (de Kroon et al., 1994; Dong and de Kroon, 1994; de Kroon and Hutchings, 1995).
Most studies on plastic alteration of clonal growth dealt with responses to variations of resource availability [light, Stuefer and Huber (1998); nutrient, Ye et al. (2006)] and abiotic stress factors (water depth, Lenssen et al. (2000); concentration of heavy metals, Koivunen et al. (2004)]. Surprisingly, thigmomorphogenetic responses (i.e. responses following mechanical stimulation; Jaffe, 1973; Telewski, 2006) due to wind, current or waves on clonal growth have rarely been investigated (e.g. see Peralta et al., 2006; Puijalon and Bornette, 2006; Kotschy and Rogers, 2008), whereas mechanical stresses are frequently encountered by plants, both in terrestrial and aquatic ecosystems.
The thigmomorphogenetic responses relating to clonal growth that have been identified are an increased number of stolons and allocation to clonal growth (Puijalon and Bornette, 2006; Liu et al., 2007), reduced stolon lengths (Doyle, 2001), altered spacer lengths [increased, Jensen and Bell (2001); reduced, Puijalon and Bornette (2006)] and tussock shape (Asaeda et al., 2005). Mechanical stress linked to current exposure could lead to alignment of creeping stems with the flow direction (Asaeda et al., 2005; Kotschy and Rogers, 2008). However, given the scarcity of data available, it is still impossible to identify general patterns on the thigmomorphogenesis for clonal growth. Consequently, the primary aim of the present study was to investigate the clonal plasticity induced by mechanical stress consisting of continuous current encountered by aquatic plants. As far as is known, only Liu et al. (2007) measured the plastic responses to mechanical stress among genetically identical ramets, whereas all other studies did not test explicitly if the morphological responses were due to phenotypic plasticity. In the present study, the clonal plasticity induced by mechanical stress was investigated by using genetically identical ramets.
It is hypothesized that mechanical stress from current velocity should induce variations of clonal traits that lead to escape from stressful conditions. This hypothesis was tested on two aquatic plant species with contrasting clonal growth forms. Such plants are an interesting model for such work because they are characterized by high levels of clonality and plasticity (Grace, 1993; van Groenendael et al., 1997; Santamaria, 2002). For each species, an experiment consisting in the transplantation of ramets of the same clone in two contrasting flow velocity conditions was carried out, and clonal traits were measured. Selected traits were: (a) relative allocation to clonal growth, (b) relative allocation to creeping stems in the clonal biomass, (c) number and (d) total length of creeping stems, (e) spacer length and (f) main creeping stem direction. Under running conditions, a reduced number of creeping stems, together with an increased length of creeping stems and of spacers, was expected. Also expected was an increased allocation to clonal growth and to spacers that promote escape from stressful conditions (Slade and Hutchings, 1987; de Kroon and Hutchings, 1995; Cain et al., 1996; Ikegami et al., 2007). As flow velocity is reduced in the downstream direction due to partial sheltering by the upstream ramets (Sand-Jensen and Mebus, 1996; Sand-Jensen and Pedersen, 1999), an increased proportion of creeping stems in a downstream direction under the stressful conditions was also expected.
The study was carried out on two aquatic plant species having contrasting clonal growth forms, Mentha aquatica L. (Lamiaceae) and Potamogeton coloratus Hornem. (Potamogetonaceae). Mentha aquatica has running stems with erect ends, completely or partly submerged, whereas P. coloratus produces stolons with regularly distributed ramets, growing totally submerged (Fig. 1). Both species are able to colonize both standing and running waters (Haslam, 1987), suggesting they are able to develop significant plastic responses to current. Throughout the study, creeping stems are referred to as stolons for P. coloratus and running stems for M. aquatica.
Submerged plants were collected from standing conditions in a groundwater-fed channel of the Ain River (France, 05°15′37″E; 45°57′57″N). Twenty-one widely spaced clones consisting of two young interconnected ramets were collected for each species. For M. aquatica, an individual ramet was defined as an anchored erect shoot. One ramet of each clone was assigned to a ‘standing’ treatment and the other one to a ‘running’ treatment. The clones were selected so that the size of the two ramets of the clone was as similar as possible. The fresh mass of the ramets was measured and did not differ significantly for the two sets of plants for both species (P = 0·29 and P = 0·61 for M. aquatica and P. coloratus, respectively; mixed models ANOVA; see Statistical analysis for details regarding the method).
All of the transplantation was carried out within 10 d in May 2003, and plants were held for no more than 24 h after collection, and before transplantation. The transplantation site was located in a man-made experimental canal, and fed with water that contained favourable nutrient levels. Ramets were planted in plastic containers (18 × 24 × 10 cm) filled to the brim with river sand (0–5 mm). All stolons and rhizomes were removed from each ramet before transplantation. For each species, three ramets were planted in separate containers (i.e. 14 containers per species), and the containers were randomly positioned within an area with standing water in the experimental canal. Planks deflecting the flow in the canal were placed around this area to ensure standing conditions. After an acclimation period of 5 weeks, for each species seven of the 14 containers were positioned in a riffle area with continuous running water located 10 m upstream the standing area (i.e. the ‘running water’ treatment). The two areas (standing and running) were characterized by similar water depth (15–20 cm above sand level). Planks deflecting the flow were placed in order to maintain permanent water renewal of the standing area: the downstream part of the standing area was opened and openings in the planks enabled water inputs on the sides of the containers. This set-up was used to reduce as much as possible potential differences in water physico-chemistry between both areas (essentially temperature and oxygen concentration).
Flow velocity encountered by plants in the running area was measured every 10 d, on eight dates, covering the whole experimentation period. Flow velocity was measured 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 estimate of average flow velocity in the water column. Flow velocity encountered by plants submitted to the ‘running treatment’ during the experimentation was 0·19 ± 0·08 ms−1, which has been demonstrated to alter plant morphology in previous studies (Power, 1996; Puijalon and Bornette, 2004). In order not to disrupt the clonal growth pattern after transplantation, the stolon and rhizome parts were allowed to grow out of the box where the mother ramet had been planted during the whole experiment.
All individuals were harvested 10–12 weeks after treatment application (in September 2003). The starting date and duration of the experiment were designed (a) to enable transplants to grow under standing versus running conditions for a period long enough to allow the turn-over of plant tissues, and (b) to harvest plants before autumnal decay. Both species displayed high survival rate (21 and 17, 19 and 19 plants survived for M. aquatica and P. coloratus in standing and running treatments, respectively). Following harvest, plants were stored in aerated tap water at 16 °C for a maximum of 2 d until measurements of clonal traits were completed.
The transplanted ramet was referred to as the ‘mother ramet’, whereas new ramets produced vegetatively by the mother ramet were referred to as ‘daughter ramets’. The following six traits were measured on creeping stems (Fig. 1).
These traits were expressed relatively to the dry mass of the mother ramet to remove a possible effect of the size of the mother ramet on the length and the number of creeping stems produced.
For measuring the dry mass, plants were divided into mother ramets, creeping stems and daughter ramets, and the different parts were weighed after drying for 48 h at 85 °C to obtain the dry mass measurements.
For all traits except creeping stem direction, the plant response to current was analysed using mixed-effects models, which enable to take into account, within the same analysis, fixed effects associated with treatment and random effect associated with clone sampled randomly in a population, and the dependence of the data measured on the two ramets of the same clone (Pinheiro and Bates, 2000). In all the models, flow condition (treatment) was considered as a fixed effect and clone as a random effect. As missing values are a problem for fitting such models, a conservative approach, which consists of omitting the incomplete clones (i.e. clones with only one surviving ramet out of the two ramets transplanted, corresponding to four and five clones for M. aquatica and P. coloratus, respectively), was used. All mass and length variables were loge-transformed prior to analysis to improve the normality of residuals and homogeneity of variances.
Number and total length of creeping stems and spacer length were analysed as an absolute value and corrected for clonal (spacer length) and mother dry mass (number and total length of creeping stems) to compare morphologies for standardized size.
Mixed-effects ANOVA were used to test (a) the absolute values of total length of creeping stems and spacer length, and (b) the differences in plant fresh mass at the transplantation date.
Mixed-effects ANCOVA were used to test the effect of treatment on (a) total length of creeping stems and spacer length corrected for size (mother ramet dry mass and clonal dry mass used as covariate, respectively) and (b) dry mass allocation, using allometric relationships (Jasienski and Bazzaz, 1999). Allocation to clonal growth within the clone was analysed using clonal dry mass as the dependant variable and total dry mass of the clone (i.e. mother dry mass + clonal dry mass) as the covariate. Allocation to creeping stems within the clonal dry mass was analysed using creeping stem dry mass as the dependant variable and clonal dry mass as the covariate.
A generalized linear mixed-effects models (GLMM) was used for the number of creeping stems produced, which was a Poisson variable. This trait was tested as an absolute value and corrected for mother ramet size, adding mother ramet dry mass as a covariate.
For all the analyses, an homoscedastic (i.e. equal variance between treatments) and an heteroscedastic model (i.e. unequal variance between treatments) were tested using likelihood ratio test, to take into account the difference in trait variance possibly induced by treatment (Pinheiro and Bates, 2000). This test was carried out on the full model with the interaction term of fixed effects (results not shown). Moreover, for analyses with a covariate, the covariate, treatment and interaction term (covariate × treatment) were first introduced in the model. Non-significant interaction terms (covariate × treatment) were eliminated from the model to obtain the final model.
The direction of the creeping stems produced was analysed using a generalized linear model, treating the number of creeping stems as the dependent variable (count variable), and the treatment and the direction as fixed effects. A clone effect was not introduced because the number of clones with data for the two ramets was too low (because some mother ramets did not produce creeping stems or because stems were too small for recording their direction). Consequently, this analysis does not take into account the dependence of some data (i.e. clone effect for clone with data for both ramets).
R-2·4·1 (R Development Core Team-2006) was used for all statistical analyses.
For both species, dry mass allocation to clonal growth did not differ between standing and running treatments (Fig. 2 and Tables 1 and and2).2). For P. coloratus, dry mass allocation to creeping stems within the clonal dry mass increased in response to current (Fig. 2 and Table 2), whereas it did not differ between treatments for M. aquatica despite a significant interaction between the covariate and flow conditions for allocation to creeping stems (see Fig. 2 for median plant size; Table 1). Only for M. aquatica, the number of creeping stems was higher under running conditions than under standing ones (2·9 and 4·1 on average for standing and running treatments, respectively), whatever the way it was considered (absolute value or relative to the mother ramet size; Fig. 2 and Tables 1 and and3).3). The total length of creeping stems was higher under running conditions when considered as an absolute value for P. coloratus and relatively to plant size for M. aquatica (Fig. 2 and Tables 1–3). For P. coloratus, spacer length significantly increased (3·0 and 5·5 cm on average for standing and running treatments, respectively; Fig. 2 and Tables 2 and and33).
For both species, the main direction of creeping stems differed between standing and running conditions (χ22 = 11·6, P = 0·003 and χ22 = 16·2, P = 0·0003 for M. aquatica and P. coloratus, respectively; Fig. 3). That is, for both species, exposure to current induced a significant increase of the proportion of creeping stems in the downstream direction (contrast test, χ21 = 5·6, P = 0·02 and χ21 = 13·4, P = 0·0002 for M. aquatica and P. coloratus, respectively; Fig. 3) to the detriment of creeping stems perpendicular to flow. The proportion of creeping stems growing in the upstream direction did not vary for M. aquatica (contrast test, χ21 = –0·69, P = 0·24) and slightly increased for P. coloratus (contrast test, χ21 = 1·79, P = 0·05).
By using pairs of genetically identical of ramets, the present study demonstrates that exposure to current flow induced significant plastic variation in clonal traits for both submerged aquatic plant species. As the observed plastic responses were clearly related to the direction of the water current, and differences in water physicochemistry were minimized by the experimental set-up, the observed plastic responses are attributed predominantly to the mechanical stress encountered by plants under running conditions. Clonal architecture was altered by exposure to current but without variations in the total allocation to clonal growth, contrary to a previous study that demonstrated for Berula erecta an increased allocation to clonal growth (Puijalon and Bornette, 2006). For P. coloratus, morphological traits that varied significantly showed the expected trend: under stressful conditions, total length of creeping stems and spacer length were higher and plants allocated more biomass to spreading organs (creeping stems). Thus, in accordance with the present hypothesis, clonal growth was altered in a sense that could favour escape from stressful conditions. However, the response observed on M. aquatica partly contradicted the hypothesis. The increased number of creeping stems observed on M. aquatica may not favour escape from a stressful habitat but rather a higher occupation of those habitats (de Kroon and Hutchings, 1995; Ikegami et al., 2007). The benefit of the increased number of creeping stems could be an enhanced anchorage efficiency and the formation of a denser canopy that might reduce effects of aero- or hydrodynamic forces (Sand-Jensen and Mebus, 1996; Speck, 2003; Liu et al., 2007).
The displayed variations of clonal architecture may represent different strategies to cope with exposure to current velocity. On the one hand, the plastic responses of P. coloratus led to an escape strategy with a guerrilla growth form, but it restricts benefits with respect to sheltering and anchorage. On the other hand, the plastic responses of M. aquatica, could lead to a resistance strategy. Indeed, as the response could result in a denser canopy possibly reducing stress on individual shoots due to sheltering and to enhancement of anchorage efficiency. These two strategies could result from a trade-off between the number and the length of the creeping stems, leading either to a higher escape (longer creeping stems) or to a higher resistance (higher number of creeping stems).
The alteration of creeping stem direction results in their grouping in the flow direction. This response can be viewed as an escape from the most stressful conditions resulting in location of ramets in more favourable conditions, benefiting potentially from the sheltering of upstream ramets (Sand-Jensen and Mebus, 1996; Sand-Jensen and Pedersen, 1999). This plastic response could have important consequences for plant maintenance in stressful habitats. Individual ramet positioning determines directly flow velocity and forces within and around plant beds. The alignment of the stems and ramets of the clones results in an overall streamlined shape at the clone scale, having a reduced frontal area exposed to flow. This spatial configuration is potentially beneficial for the whole clone through an enhanced growth of downstream ramets that are more sheltered. This capacity to expand clonally under running conditions may be essential for maintenance and spreading, as sexual reproduction is frequently reduced under such conditions (Pollux et al., 2006).
In clonal plants, the concept of division of labour is generally related to specialization, and co-operation based on vascular conductivity of stolon internodes (Stuefer et al., 1998). The present study suggests that, in addition to the transfer of resources, attenuation of mechanical stresses may, at least in hydrodynamic environments, be an equally important aspect of the division of labour for the whole clone performance. This is in agreement with earlier findings for the intertidal salt-marsh species Spartina, where the most exposed shoots attenuate the majority of the hydrodynamic energy, thereby sheltering the rest of the clone and enhancing accretion of sediment and nutrients within the tussock (Bouma et al., 2005).
The alignment of creeping stems with the flow direction may be induced by different mechanisms. First, it could be a passive consequence of the hydrodynamic forces exerted on growing stems (drag forces pushing the stems in the flow direction; Vogel, 1994; Kotschy and Rogers, 2008). Secondly, it could also result from a higher mortality of the creeping stems growing in the other directions, for instance due to damages by sandblasting and drifting particles (Cleugh et al., 1998). Thirdly, the alteration of creeping stem direction may be the result of a differential activity of the meristems. The meristems could grow preferably in the downstream direction that is partly sheltered from current and submitted to less stressful conditions (Sand- Jensen and Mebus, 1996; Sand-Jensen and Pedersen, 1999), whereas other growth directions could be inhibited. Unravelling the underlying mechanisms remains a topic for further study.
In conclusion, in the present study, it was demonstrated experimentally that plastic variation in clonal growth occurs in response to flow stress and might lead to adaptation to these stress conditions. For the same relative allocation to clonal growth, species display responses that could result either in escape from stressful conditions through formation of a guerrilla growth form or in enhanced resistance through formation of a denser canopy. Consequently, further studies on the capacity of clonal plants to adapt to mechanical stress should take into account both the morphological plastic responses at the ramet scale and the clonal plastic response.
We thank P. Robinot (Poissons Sauvages Production) for providing the experimental canal and D. Reynaud and E. Malet for technical assistance. This study was partly funded by the ‘Cluster Environnement’ of the Rhône-Alpes Region and was carried out under the aegis of the long-term ecological research programme on the Rhône River Basin (Zone Atelier Bassin du Rhône).