Morphological variations of natural populations of an aquatic macrophyte Elodea nuttallii in their native and in their introduced ranges

Introduced plant species appear often more vigorous than their conspecifics in the native range. We investigated the morphological variations of Elodea nuttallii (Planch.) St John and its behaviour in its native and in its introduced habitats. We assessed eight morphological traits of 24 populations in the native North America range and 16 populations in the introduced European range. The introduced taxa can be very distinct in their growth form and size from counterparts in their native range. The shorter broaded-leaved phenotype typically occurs in shallow streams, whereas the longer spacer narrow-leaved phenotype occurs in lakes. Larger leaf width and higher number of lateral shoots when nutrients are not limitingmay enhance plant performance. The European populations grow more vigorously than their American relatives, possibly because of different selection pressures.


Introduction
Several factors that may influence the susceptibility of a community to invasion have been identified (Crawley 1987;Lonsdale 1999;Davis et al. 2000).The reasons that some species become invasive following introduction, and which traits allow these species to disrupt the ecosystems they invade, have become topics of considerable interest and debate (Blossey and Nötzold 1995;Callaway and Aschehoug 2000;Keane and Crawley 2002;Müller-Schärer et al. 2004).While all non-native plants may carry some risk of becoming natural invaders, the taxonomic analysis of Daehler (1998) suggests that plants with amongst the highest risk of becoming natural invaders worldwide include species that are primarily aquatic or semi-aquatic.Furthermore, invaders characterized by higher leaf area and greater phenotypic plasticity appear to be "super invaders" (Daehler 2003).For example, the variation of populations of Elodea nuttallii (Planch.)St John, a well known invasive species, is such that the two extremes of its morphological range appear to be distinct taxa in Europe, one with long, completely plane leaves, and the other with shorter, broader, strongly reflexed leaves (Simpson 1988;Vanderpoorten et al. 2000).
Understanding the basis for invasiveness is critical in assessing the risks associated with new or spreading invasive species.Apart two flora comparisons (Crawley 1987;Thébaud and Simberloff 2001) the available literature does not provide sufficient empirical data to suggest that increased vitality in invasive populations is a common phenomenon.A previous work, based on literature data, established that there is no general tendency for aquatic plants to be more vigorous in their introduced ranges (Thiébaut 2007).The aim of our study was to collect empirical data for one of the most rapid current invader of Europe in water bodies, Elodea nuttallii.Native to North America, E. nuttallii was introduced into Europe in 1939 (Simpson 1984).Only female plants were observed initially, introduced via the trade in live aquarium plants (Cook and Urmi-König 1985).The spread of E. nuttallii has resulted in its displacement of E. canadensis Michaux from many localities where the latter had previously become well established in Europe (Simpson 1990;Thiébaut et al. 1997;Barrat-Segretain 2001;Larson 2007).In the long term, E. canadensis and E. nuttallii therefore appears redundant enough to coexist within the sites (Hérault et al. 2008).The different invasiveness patterns of Elodea species could be solely due to the ecological drift (Hérault et al. 2008).In North America, E. nuttallii is concentrated around the St Lawrence Valley, the Great Lakes, and the Pacific West Coast, but is more common further south.E. nuttallii populations are rarely troublesome in natural habitats in North America, but plants can become dominant in altered or created aquatic systems, especially when bicarbonate, reduced iron, and phosphorus are plentiful.E. nuttallii is listed in southern Ontario (Argus and White 1977), southern British Columbia and southern Quebec (Bouchard et al. 1983).Remarkably little is known about the ecology of E. nuttallii within its native range.Studies indicate that E. nuttallii is present at different stages in the eutrophication of a lake (Lawrence 1976).
The distribution of the aquatic plants in North America and in Europe has been documented in the literature and by consulting herbariums in Museum in Canada (Montreal, Quebec) and in the USA (Albany, New York State).Even though the climate was generally more continental in North America, all regions were characterized by a temperate climate.For a representative comparison of North American and European regions, a total of 90 populations were visited during the summers of 2003 and 2004.The populations were distributed through four main regions in both continents: thirty European populations were located in Britain, northeastern France, Switzerland and Belgium, while 60 North American populations were located in Canada and the United States (30 sites in Quebec and 30 in New York State; Annex 1).We investigated field morphological variations of natural populations of E. nuttallii in their native (North America) and in their introduced ranges (Europe).

Morphological variation of E. nuttallii
To measure the morphological variation of Elodea nuttallii in its native ranges and in its introduced ranges, 16 sites were selected in Europe from among the 30 European sites visited, while of the 60 North American sites visited, 11 and 13 sites, respectively, were chosen in Canada (Quebec) and the USA (New York State).Among them, five and ten of the sites in Canada and USA respectively were lakes whilst eight of the sites in Europe were also lakes.Within each region, the populations were sampled according to a geographical gradient (samples were distributed evenly in the East-West and South-North areas of each sampling region).Records made for each population included geographical location, the most frequent species in the accompanying flora and the type of water body (lake or stream).The criteria used to identify the Elodea species in North America and in Europe were the criteria defined by Simpson (1988).Field trips were carried out in the summer (from mid-July to mid-August).Thirty plants were randomly collected at each site to investigate the morphological variations.Measurements of morphological traits were realized in the laboratory.The presence or absence of fruits in each plant was investigated.The main traits are illustrated in the Figure 1.
Eight morphological traits were measured on each plant: (1) main shoot length; (2) number of lateral shoots -Lateral shoots included the initial lateral shoot which developed from the nodes on the original apical shoot and the other lateral shoots that developed either from the same nodes or from the nodes on the lateral shoots -; (3) cumulated length of lateral shoots (initial +secondary); (4) bulk, measured as the ratio between dry mass of the plant and total length of all shoots; (5) internode length, defined as the length of ten internodes below the three-cm long apex; (6) length of a leaf located at the sixth whorl; (7) width of a leaf located at the sixth whorl; (8) area of a leaf located at the sixth whorl.The leaf fully expanded was fixed on a paper, and its length, its width and its area were calculated by using a Scion image V. 1.63 programme).
Only sites sampling several times were selected to represent base flow chemistry.Twelve sites (five rivers and seven lakes) and thirteen sites (six streams and seven lakes) were respectively chosen in North America and in Europe.Conductivity and pH were measured using a combined glass electrode.Soluble reactive phosphorus (SRP) was determined using the ascorbic acid method for SRP (APHA 1992), while ammonium was analyzed using the phenol hypochloride method (Wetzel and Likens 1991).Nitrate, sulfate and chloride analyses were determined with ion chromatography.
For lakes, we added two parameters: water clarity and Chlorophyll a. Water clarity was estimated by Secchi disk according to standard methods.Chlorophyll a is measured by filtering a sample water through a glass fiber filter.The filter is ground up in an acetone solution and Chlorophyll a quantified using a spectrophotometer.

Statistical analyses
We used analysis of variance (two-factor ANOVAs) methods for testing for differences in plants traits between continents and type of water body.Data were log transformed to normalize their distribution prior to analysis.Tukey HDS tests were used as a post-hoc test.
They were set at a significance level of 0.05.
Spearman's coefficients of correlation were calculated between chemical variables and independent morphological traits.Relationships with the traits "leaf area" and "bulk" were not considered because of their correlations with other traits.
Canonical Correspondence Analysis (CCA) based on raw data was used to elucidate the relationships between morphological traits (except the area leaf and bulk) and environmental variables.Significance of the canonical axes was tested using Monte-Carlo permutations of samples.

Results
The species E. nuttallii typically form dense monospecific stands in the lower littoral zone of lakes in the prospected sites in Canada, whereas it was often mixed with Elodea canadensis (38% of the sites) in eastern USA.The two Elodea species were only found together in 6.7 % of the European sites.In the introduced range, E. nuttallii often covered the littoral zone of the shallow basin or streams with dense mats, but it is most commonly associated with European native aquatic vascular plants (e.g.Callitriche species, Ranunculus species, Myriophyllum spicatum L.).

Morphological variation of E. nuttallii in Europe and in North America
All morphological traits significantly differed between Europe and North America and between lakes and streams (Tables 1, 2).The average population size (main shoot length, internode length), bulk, branching (number and length) and leaf traits (area, length, and width) were significantly higher in European than in American populations (Table 1).A highly significant difference was established between streams and lakes, except for main shoot length and length leaf (traits 1, 6, Table 2).The interactions between the continent and the type of waterway were significantly different for the morphological traits tested, with the exception of length of lateral shoots (two-factor ANOVAs, Table 3).

Linking species and environmental variables
The number of lateral shoots, bulk and width leaf were mainly correlated with nutrients (traits 2, 7), whereas the main shoot length, length of lateral shoots and the internode length were correlated with the Secchi Disk (traits 1, 3, 5; Table 4).
The CCA showed that a high percentage of the traits-environment relation (96.7%) was explained by the first two canonical axes (Figure 2).Monte Carlo permutation tests on the first eigenvalues indicated that the relation between the traits and the environmental variables selected by the model were significant.These results suggest that morphological traits of Elodea nuttallii in the European and North American sites can be explained by the variables included in the model.The first axis was defined positively by leaf width (trait 7) and in lesser extend by number of lateral shoots (trait 2).The second axis was characterised negatively by main shoot length (trait 1), in lesser extend by the length of leaf (trait 6) and by internode length (trait 5).A clear opposition was established between native population traits and introduced population traits (Figure 2).Three European populations corresponding on recently introduced species are close to North American populations.

Morphological variation of E. nuttallii in Europe and in North America
Despite considerable variation between populations within continents, there were pronounced differences between continents; the vitality (main shoot length, branching,) was significantly higher in European than in American populations.Introduced plant species that became successful invaders appear often more vigorous and taller than their counterparts in the native range (Thebaut and Simberloff 2001;Leger and Rice 2003;Stastny et al. 2005).This seems to contradict the results of Daehler (2003) who reviewed the performance of cooccurring native and alien species but did not find higher growth rates, competitive ability nor fecundity to be characteristics of the latter.The seemingly contradictory conclusions of Daehler's review (2003) could be due to fact that his review was not confined to congeners (Pysek and Richardson 2007).
The introduced Elodea nuttallii populations appeared to perform better in Europe, possibly because of changed selection pressures (no selective herbivores for example).Johnson et al. (1998) suggested that Acentria ephemerella , an European aquatic moth larva, can potentially feed on Elodea species in North America.However when the density of these grazers is high, herbivory by Acentria ephemerella causes severe damage to the European species Myriophyllum spicatum.Moreover, Lake and Leishman (2004) showed that both invasive exotic and non-invasive exotic species had significantly lower levels of leaf herbivory than native species, implying that release from pests alone cannot account for the success of invasive species.This is congruent with our results.We observed no leaf damage in the European populations, whereas limited damages were observed in the leaves of few American populations (6.7 % of the sites).Elodea nuttallii suffer less herbivory than native species in Europe (Thiébaut and Gierlinski 2007).The vitality of invasive plants may be associated with changes in resistance as well as tolerance to herbivory and that both types of anti-herbivore defence may need to be examined simultaneously to advance our understanding of invasiveness.

Biological traits of E. nuttallii in native and in introduced areas
There seem to be some traits commonly associated with successful invaders: growth rate, resistance and tolerance to disturbances, high fecundity and efficient dispersal of seeds or vegetative fragments.
North American and European populations of E. nuttallii showed pronounced differences in reproduction.E. nuttallii was characterized by no sexual reproduction in their introduced range, whereas fruits were observed in 6.7 % of the USA sites.Fruiting specimens are very rare in North America (Lawrence 1976) and very few fully mature fruits were noticed in Canada (Catling and Wojtas 1986).Although E. nuttallii reproduces both sexually and asexually by vegetative clonal propagation in North America, vegetative reproduction seems to be the dominant method of propagation.To our knowledge, no study has been undertaken to investigate the speed and extent of vegetative reproduction and propagation of E. nuttalii in North America.In the introduced range, Barrat-Segretain et al. (2002) showed that when introduced to a new habitat, the establishment of Elodea buds is rapid, since the propagules sank into the sediment and grew rapidly.However, it is difficult to drawn any conclusion regardless the role of the sexual versus vegetative reproduction on the propagation of E. nuttallii in their native and introduced ranges.
The clearest consistent difference that Lake and Leisham (2004) found between invasive exotic and non-invasive exotic species was in specific leaf area, suggesting that large specific leaf area facilitates invasiveness.Due to infrequent and ephemeral flowering, together with the minute size of the flowers, leaf morphology is often the trait used to differentiate Elodea species in North America (Catling and Wojtas 1986;Lawrence 1976).Unfortunately there is overlap in the characters and there is little agreement among authors on the nature of these characters.According to Catling and Wojtas (1986) leaf width partially separates male plants of Elodea canadensis and Elodea nuttallii but not female plants.Determination of sterile Elodea species has also been an area of controversy in Europe.The European populations of E. nuttallii have a higher specific leaf length (1.17 cm) and leaf width (0.22 cm) and internode length (0.74 cm) than in native ones (0.96 cm, 0.15 cm, 0.56 cm respectively).The leaves are either linear or linear-lanceolate and the shape of the leaf apex is narrowly acute to acuminate in E. nuttallii's European populations.By changing leaf width, individuals maximize growth under a variety of environmental conditions and have the potential to intensify the development of lateral shoots and of resource-acquiring structures in the more favourable microhabitats of a heterogeneous environment.Larger leaf width and higher number of lateral shoots were produced in streams (Table 3) because there is a stimulating effect of moderate flow, a response to flow constraints or simply as a function of greater nutrient supply in rivers (mean values: ([ammonium] = 32µg/l; [SRP] = 51µg/l) than in lakes ([ammonium] = 48µg/l; [SRP] = 16µg/l).
Invaders seems often have higher phenotypic plasticity than natives, and this plasticity probably allows invaders to succeed in a wider range of environments (Daehler 2003).A greater phenotypic plasticity is particularly advantageous in disturbed environments.Plastic "general-purpose genotypes" could have a fitness advantage in founding populations where local adaptation has not yet occurred (Sexton et al. 2002), or cannot occur because of a lack of genetic variation.Genetic differentiation in introduced populations may occur in any ecological trait that is beneficial under the novel selection conditions, given that there is genetic variation for it.Our results established that E. nuttallii exposed to environmental stresses showed a wide range of morphological variations.
The shorter broaded-leaved phenotype typically occurs in shallow streams, whereas the longer spacer narrow-leaved phenotype occurs in deeper waters.The latter may also be present near the surface when light intensity is reduced due to turbidity of the water.High phenotypic plasticity is really a key ingredient of invasiveness of plants in aquatic environments.For example, Di Nino et al. (2007) showed that genetically uniform populations of E. nuttallii develop a common phenotype under similar environmental conditions.However, with increasing water nutrient enrichment, E. nuttallii showed increases in leaf area with decreases in internode length (Di Nino et al. 2007).These plastic responses may enhance plant performance.To detect evolutionary change in invaders, comparative studies of native versus introduced populations are needed.A combination of field and common garden studies is needed to fully understand evolutionary change in E. nuttallii.

Implication for management
On the basis of our results (a high plasticity, allocation of resources to growth and a foliar investment), E. nuttallii should be regarded as having a high risk of being invasive, and we recommend that it is nominated as a priority target for eradication or control in new sites.Different studies established that E. nuttallii is probably in an expansion phase in Europe and will likely spread to new areas (Simpson 1984;Thiébaut et al. 1997;Barrat-Segretain 2001, Larson 2007).In general, the priority of biological invasion control is to prevent new infestations from taking hold, especially for the fastest-growing and most disruptive species.It should be noted that most stakeholders and decision makers have a limited perception of the problem (Garcia-Llorente et al. 2008).The prevention approach would include greater investment in prevention: preventing organisms from entering a particular pathway, and preventing organisms that are transported from being released or escaping alive.Stronger enforcement of existing laws coupled with an intensive public education campaign is needed to prevent further Elodea species introduction.
Conditions where invaders had the largest performance advantages (high or fluctuating nutrient loading, flow regime alteration, industrial contamination) are generally associated with degraded habitats.Biological invasions could be seen as one of a syndrome of traits that characterize degraded aquatic ecosystems (Wilby 2007).The European Water Framework Directive requires the restoration of degraded freshwaters (Wilby 2007).A basic recommendation is to foster good ecological conditions in water masses as a self-defence measure for aquatic ecosystems against the colonising pressure of Elodea species.Specific emphasis on hydromorphological as well as nutrient-related pressures, combined with the delivery of programmes on a catchment scale, should support the environmental improvements needed to sustain reduction of Elodea species.

Figure 1 .
Figure 1.Diagram illustrating E. nuttallii and its morphological traits

Figure 2 .
Figure 2. Ordinations obtained by Canonical Correspondence Analysis based on morphological traits.The relative position and magnitude of the biological traits are represented by arrows.The sites are positioned as points in the C.C.A. diagram.Trait 1: main shoot length; trait 2: number of lateral shoots; trait 3: length of lateral shoots; trait 5: internode length; trait 6: leaf length; trait 7: leaf width

Table 1 .
Morphological differences between European and American populations of Elodea nuttallii P-values (Mean± Standard deviation)

Table 2 .
Morphological differences between populations of Elodea nuttallii in lakes and in streams P-values (Mean± Standard deviation, ns: not significant)

Table 3 .
Two-factor ANOVAs used to test the continent and water bodies effects on each trait (ns: not significant)

Table 4 .
Coefficient of correlation between water chemical parameters and plant traits of Elodea nuttallii