Ecophysiological differences between genetic lineages facilitate the invasion of non-native Phragmites australis in North American Atlantic coast wetlands


  • Thomas J. Mozdzer,

    Corresponding author
      Correspondence author. E-mail:
    Search for more papers by this author
    • Present address: Smithsonian Environmental Research Center, PO Box 28, Edgewater, MD 21037, USA.

  • Joseph C. Zieman

    1. Department of Environmental Sciences, University of Virginia, 291 McCormick Rd, PO Box 400123, Charlottesville, VA 22904, USA
    Search for more papers by this author

Correspondence author. E-mail:


1. Over the last century, native Phragmites australis lineages have been almost completely replaced along the North American Atlantic coast by an aggressive lineage originating from Eurasia. Understanding the mechanisms that facilitate biological invasions is critical to better understand what makes an invasive species successful.

2. Our objective was to determine what makes the introduced lineage so successful in the study area by specifically investigating if morphological and ecophysiological differences exist between native and introduced genetic lineages of P. australis. We hypothesized a priori that due to phenotypic differences and differences in plant nitrogen (N) content between lineages, the introduced lineage would have a greater photosynthetic potential.

3.In situ ecophysiological and morphological data were collected for 2 years in a mid-Atlantic tidal marsh and in a glasshouse experiment. We measured photosynthetic parameters (Amax, water use efficiency, stomatal conductance) using infrared gas analysis, in conjunction with ecophysiological and morphological parameters [specific leaf area (SLA), leaf area, chlorophyll content, N content].

4. Introduced P. australis maintained 51% greater rates of photosynthesis and up to 100% greater rates of stomatal conductance which are magnified by its 38–83% greater photosynthetic canopy compared to the native type. The introduced lineage also had a significantly greater SLA and N content. Glasshouse-grown plants and naturally occurring populations demonstrated similar trends in ecophysiological characteristics, verifying the heritability of these differences. These ecophysiological differences, when combined with an extended growing season, provide the mechanism to explain the success of introduced P. australis in North America.

5. Our findings suggest the native type is a low-nutrient specialist, with a more efficient photosynthetic mechanisms and lower N demand, whereas the introduced type requires nearly four times more N than the native type to be an effective competitor.

6.Synthesis. Our study is the first to combine field and laboratory data to explain a biological invasion attributed to ecophysiological differences between genetic lineages. Our data corroborates earlier work suggesting anthropogenic modification of wetland environments has provided the state change necessary for the success of introduced P. australis. Finally, our results suggest that genotypic differences within species merit further investigations, especially when related to biological invasions.


Invasive species are a significant problem for both terrestrial and aquatic ecosystems due to the potential changes they can cause in ecosystem services and ecosystem function (Bertness, Ewanchuk & Silliman 2002). The literature is full of plant invasion examples demonstrating how introduced congener and exotic invaders are often more successful than native plants physiologically (Schweitzer & Larson 1999; McDowell 2002; Deng et al. 2004), phenologically (Brock, Weinig & Galen 2005) or due to enemy release (Keane & Crawley 2002; DeWalt, Denslow & Ickes 2004; Meyer & Hull-Sanders 2008). In Pacific coast ecosystems, there are also examples of intrageneric invasions (Daehler & Strong 1996), as well as intraspecific hybridization events resulting in hybrids with an even greater vigour (Anttila et al. 2000). However, multiple introductions and/or hybridization events have not always resulted in a physiologically superior genotype (Brodersen, Lavergne & Molofsky 2008). While hybridization among genotypes increased phenotypic plasticity in Phalaris arundinacea, ecophysiological differences among populations were not found to be important with respect to invasion potential when compared with morphology and secondary defence compounds (Brodersen, Lavergne & Molofsky 2008). Less understood is how conspecific genetic lineages can also act as invasive species, completely displacing native genetic lineages.

The common reed, Phragmites australis (hereafter Phragmites) is one of the world’s most successful emergent macrophytes. Described by some to be the world’s most widely distributed angiosperm (Holm et al. 1977; Mal & Narine 2004), Phragmites has been a component of North American wetlands for at least 10 000 years (Niering, Warren & Weymouth 1977; Orson 1999). However, over the last century, Phragmites has invaded tidal and non-tidal wetlands, displacing native vegetation (Chambers, Meyerson & Saltonstall 1999), resulting in a fivefold reduction in species richness (Bertness, Ewanchuk & Silliman 2002), reductions in benthic faunal habitat use (Meyerson et al. 2000; Osgood et al. 2003) and changes in ecosystem function (Meyerson, Vogt & Chambers 2000; Windham 2001). The recent invasive nature of Phragmites has been attributed to the cryptic invasion of a non-native (Eurasian) lineage introduced sometime in the 1800s (Saltonstall 2002). Along the North American Atlantic coast, the native lineage, haplotype F, had the widest historical distribution; herbarium studies have demonstrated its replacement by the introduced lineage, haplotype M, over the past century (Saltonstall 2002). Although non-native Phragmites has been a component of North American wetlands for over a century and hybridization is possible (Meyerson, Viola & Brown 2009), no naturally occurring hybrid populations have been found (Saltonstall 2003). The lack of conspecific hybridization presents a unique opportunity to evaluate potential differences in ecophysiological traits between the native lineages and the conspecific introduced lineage since these traits may be highly conserved (Brodersen, Lavergne & Molofsky 2008).

Saltonstall’s (2002) study demonstrated that in the course of less than a century there was an almost complete shift to introduced Phragmites throughout wetlands in the north-eastern USA. This introduced lineage, haplotype M (hereafter denoted introduced), has both expanded the range of historically occupied habitats and outcompeted native lineages, resulting in their extinction. Studies have correlated eutrophication (Chambers, Meyerson & Saltonstall 1999; Bertness, Ewanchuk & Silliman 2002), potential allelopathic mechanisms (Rudrappa, Bonsall & Bais 2007) and disturbance (Minchinton & Bertness 2003) with its expansion. However, no study has explained why one lineage is invasive and the other is not. Ecophysiological differences may help explain the observed differences in distribution and invasive potential.

Despite these recent and drastic changes in North American wetlands, relatively few studies have been published that have specifically investigated differences between these lineages and, more importantly, the mechanisms by which the aggressive Eurasian lineage is so successful when compared to the non-invasive native. The limited number of studies have shown that the introduced lineage has a higher stem density (League et al. 2006; Saltonstall & Stevenson 2007), greater foliar nitrogen (N) content (Packett & Chambers 2006; Saltonstall & Stevenson 2007; Mozdzer, Zieman & McGlathery in press), and greater N uptake rates (Mozdzer, Zieman & McGlathery in press). These studies presented results of the invasion process, and Mozdzer, Zieman & McGlathery (in press) reported that greater N uptake rates in introduced lineages may provide a mechanism belowground by which the introduced type is so successful. In this study, we investigated differences in physiological parameters related to phenotypic differences between the native and introduced lineages that dominate the Atlantic Coast (Blossey 2002). Because the phenotypic differences are relatively constant throughout North American wetlands (Blossey 2002), we examined ecophysiological differences between the two haplotypes, and how these differences may relate to invasive potential.

Here, we present the results of investigations designed to reveal intraspecific physiological and morphological differences between native and introduced lineages of Phragmites in North American tidal wetlands in Virginia and Maryland, USA. We hypothesized a priori that ecophysiological differences exist between native and introduced Phragmites. Due to differences in leaf colour (indicative of chlorophyll content) and tissue N content, we hypothesized a priori that the introduced lineage would have a greater photosynthetic potential because these characteristics have been positively correlated with photosynthetic rates (Reich, Ellsworth & Uhl 1995; Reich et al. 1995). Descriptive and ecophysiological data were collected on native (type F) and introduced (type M) Phragmites in a tidal marsh, complemented with similar data for native and introduced Phragmites grown in a common environment. By comparing plants that were growing side by side in the field, and plants grown in the glasshouse, we are able to make direct comparisons about genetically inherited differences between the native and introduced genetic lineages, which we believe can explain the success of the introduced lineage in populations along the North American Atlantic coast.

Materials and methods

Location and plant populations

Field measurements were taken between May 2005 and August 2006 in the King’s Creek tidal marsh on the Choptank River near Easton, MD (38°46′ N, 75°58′ W). The King’s Creek marsh is part of The Nature Conservancy’s Choptank Wetlands preserve, comprising 225 acres of undisturbed tidal marsh. This site was chosen since genetically identified native (type F) and introduced (type M) Phragmites populations (K. Saltonstall, pers. comm.) growing in adjacent stands allowed us to make direct comparisons between the two lineages. Porewater ammonium and phosphate concentrations (= 38), the primary limiting nutrients to Phragmites, were indistinguishable between plant communities at this site (data not presented). Therefore, we assume that any differences between plant lineages are due to heritable genetic differences alone.

Measurements were made on native (type F) and introduced (type M) plants growing in the University of Virginia glasshouse during 2004 and 2005. Native Phragmites plants were collected from an oligohaline tidal marsh on Occupacia Creek on the Rappahannock River, near Chance, VA (38°04′ N, 76°56′ W). Introduced Phragmites originated from a tidal salt marsh on the Eastern Shore of Virginia within the Virginia Coast Reserve – Long Term Ecological Research complex near Oyster, VA (37°17′ N, 75°55′ W). Both populations were chosen because they were genetically identified in Saltonstall’s (2002) study. Each year, rhizome fragments and emergent shoots from these populations were cleaned of dead organic material and potted individually in 10 cm square pots in clean sand. An electronic moisture control system was used to keep the sand at field capacity. Plants were fertilized twice a week with 300 p.p.m. Peters® 20-20-20 fertilizer amended with 100 p.p.m. organic N.

To assess plant nutrient status at each of the field sites, 30 leaves of each genotype of randomly selected plants were harvested from the third node from the top of the plant to be analysed for elemental C and N content in July 2005. As introduced Phragmites does not grow adjacent to the native stand on Occupacia Creek, leaves were collected from an introduced population (identified morphologically) on Hoskins Creek (37°54′ N, 76°51′ W) for comparison of plant tissue nutrient status.

Descriptive data

To quantify morphological and ecophysiological differences between the native and introduced lineages, leaves were collected from thirty haphazardly selected plants in June, July and August 2005 and 2006 at the King’s Creek field site. Harvested leaves from the third node from the top of the plant were immediately frozen in the field at −80 °C. Leaf area (LA) was measured using a LiCor LI-3000 leaf area meter (Li-Cor, Lincoln, NE, USA) prior to freeze drying, and dry masses were recorded. Chlorophyll was extracted from freeze-dried tissue after homogenization and acetone extraction with spectrophotometric pigment determination using the equations of Lichtenthaler (1987) (= 28–32 per lineage per month in 2006). Specific leaf area (SLA) was determined on leaves by normalizing LA to the freeze-dried mass of each individual leaf (= 30 per lineage per month). Total LA per ramet was determined for 10 randomly selected ramets in July and August 2005 by measuring the LA of each leaf on an individual ramet using the LI-3000 leaf area meter (Li-Cor). In the glasshouse study, chlorophyll concentrations were measured as described above on freeze-dried tissue samples once in July 2004. Leaf tissue was analysed for total carbon (C) and N content in both field-grown and glasshouse-grown plants using an elemental analyser (NA 2500; CE Instruments, Milan, Italy).

In July 2005, stem density and biomass measurements were made on native and introduced Phragmites using 0.25 m2 quadrats at all collection sites described above. Three haphazardly placed quadrats were taken per site to estimate ambient density and biomass in introduced Phragmites monocultures. Since native Phragmites populations in general do not form monocultures like the introduced variety and are patchy in distribution (Blossey 2002; Saltonstall, Peterson & Soreng 2003), we selected areas representative of ambient native Phragmites densities.

Photosynthetic measurements

Infrared gas analysis field measurements were conducted monthly in June, July and August of 2005 and 2006 using a Li-Cor 6400 Portable Photosynthesis System (Li-Cor) at the King’s Creek site. Lab measurements were conducted on glasshouse-grown plants in 2004 and 2005 for the month of June only. Measurements were only made on clear days during the daylight hours of 10:00 to 15:00. Light response curves (LRC), measuring the change in net photosynthesis as a response to changes in photosynthetic photon flux density (PPFD), were measured each month on the third fully extended leaf on randomly selected plants for each lineage. LRC measurements were alternated between native and introduced plants throughout the day to minimize potential sampling artefacts. For all LRCs, CO2 concentration in the reference chamber was set to 400 p.p.m. and the leaf block was set to 30 °C. We measured the response of photosynthesis to eight differing PPFD levels (1500, 1000, 700, 400, 200, 100, 50 and 0 μmols s−1 m−2) using the Li-Cor 6400’s internal red + blue light source. Leaves were allowed to acclimate for at least three minutes before steady-state gas exchange properties were observed, logged and changed to the next light level using the light curve program. Data from 2005 and 2006 were pooled by month prior to modelling, and measured photosynthetic rates were fit to the Mitscherlich function Amax(1−e−Aqe(PPFD-LCP)) with PPFD, light compensation point (LCP) and Aqe (photosynthetic quantum efficiency) (Potvin, Lechowicz & Tardif 1990) for each response curve as described by Peek et al. (2002). Model outputs for each LRC were then analysed in a 2 × 3 factorial design, to investigate interactions of plant type and time of year (seasonality) in sas (version 9.1; Cary, NC, USA). Significant differences between parameter estimates were tested with Tukey’s HSD post hoc test. Because each LRC was measured on a randomly selected plant within the clonal stand, each parameter estimate was considered to be an independent measurement. Dark respiration (Rd) and the light saturating value (Is) were determined using the photosyn assist software (Dundee Scientific, Dundee, UK). In the glasshouse-grown plants, conditions were identical, except the leaf block temperature was set to 25 °C. Parameter estimates were analysed in a one-way anova with sampling date as a block.


Mean plant density, biomass, %C and %N were greater in introduced Phragmites stands than in native Phragmites in both the field collection site and the site of in situ field measurements (Table 1). Number of leaves varied significantly (F3,36 = 37.56, P < 0.0001), with the introduced type having more leaves per ramet (plant effect F1,36 = 93.3, P < 0.0001) in both July (13.8 ± 0.5 vs. 10.8 ± 0.7) and August (month effect F1,36 = 7.14, P = 0.011) (16.8 ± 0.3 vs. 10.4 ± 0.5), and a greater SLA throughout the foliar canopy than native Phragmites (124.5 ± 1.0 vs. 108.4 ± 0.7 cm2 g−1) (F1,208 = 148.3, P < 0.0001; Fig. 1). Mean LA per individual ramet was significantly greater for the introduced lineage (plant effect F1, 33 = 30.93, P < 0.0001) in both July (1210 ± 98 vs. 874 ± 91 cm2) and August (1159 ± 70 cm2; Fig. 2). However, no significant temporal effect was observed (month effect F1, 36 = 3.61, P = 0.065). Leaf area per leaf position did not vary significantly and SLA did not change significantly throughout the leaf canopy for either lineage (F15, 122 = 0.3, P = 0.99 introduced, F14, 93 = 1.47, P = 0.13 native), (Fig. 1). SLA on the third fully expanded leaf was significantly greater in the introduced lineage (125.9 ± 1.4 vs. 110.1 ± 1.0 cm2 g−1) (F1, 179 = 80.26, P < 0.0001; Fig. 1) and did not vary throughout the growing season for either plant lineage (F2, 179 = 0.04, P = 0.95).

Table 1.   Mean plant density, biomass and elemental C and N at site of field experiment and location of plant collection for glasshouse-grown plants. ‘Rappahannock introduced’ is a nearby population of introduced Phragmites on Hoskins Creek on the Rappahannock River, presented here to illustrate differences in elemental composition between native and introduced lineages within sites. Biomass and density are estimated from 0.25 m2 quadrats (= 3), and elemental C and N are estimated from individual leaves (= 30 for Choptank and = 10 for Rappahannock and VCR). N/A indicates that data are not available for this variable
 Stem density (m−2)Live biomass (g m−2)%N%C
Choptank native37.3 ± 5.8707 ± 1071.89 ± 0.0644.30 ± 0.08
Choptank introduced72.0 ± 15.12065 ± 2022.41 ± 0.0745.06 ± 0.14
Rapahannock native37.2 ± 12.0517 ± 762.16 ± 0.0745.82 ± 0.30
Rapahannock introducedN/AN/A2.70 ± 0.0846.20 ± 0.27
VCR introduced74.6 ± 13.91339 ± 2273.03 ± 0.0947.14 ± 0.16
Figure 1.

 Mean specific leaf area (SLA) ± SE and mean individual leaf area ± SE throughout the plant canopy, beginning with the first fully extended leaf for native and introduced Phragmites in July 2005. SLA did not vary significantly through canopy, but introduced Phragmites had a significantly greater mean SLA (P < 0.0001).

Figure 2.

 Mean leaf area per ramet for native and introduced Phragmites plants in Choptank wetlands field site. = 10 plants type−1 month−1.

Foliar chlorophylla + b concentrations in introduced Phragmites were consistently at least twice as high as native plant concentrations (plant effect F1, 176 = 7.42, P = 0.007), and varied throughout the growing season (month effect F2, 176 = 4.11, P = 0.018) (Table 2). There was a significant interaction with plant type and month (interaction plant × month F2, 176 = 7.71, P = 0.006), corresponding to the seasonal changes in photosynthetic rate. While the photosynthetic rate peaked in July for both lineages, seasonal differences were not significant in the introduced lineage. No significant differences were observed in the ratio of chlorophylla:chlorophyllb (P > 0.05). Peak chlorophyll concentrations were observed in July for both native and introduced lineages, corresponding to the seasonal peak in Amax (Table 1). Although native plants did show a peak value in July, post hoc tests did not indicate significant differences in native chlorophyll concentration throughout the season. Introduced Phragmites chlorophyll concentrations did vary significantly, with a peak concentration in July (Table 2).

Table 2.   Photosynthetic parameter estimates for native and introduced Phragmites in the Choptank wetlands field site (= 62 total light response curves, with = 8–11 per lineage per month, = 180 total leaf chlorophyll analyses with = 28–32 per month in 2006). Only Amax (μmol m−2 s−1) (P < 0.0001), Amax (nmol g−1 s−1) (P < 0.0001) and chlorophyll content (mg g−1 dry wt.) (P < 0.0001) varied significantly, and parameter estimates were always significantly greater in the introduced lineage within each month. Significant differences between parameter estimates are indicated by different letters. Neither quantum efficiency, Aqe, dark respiration, Rd, the light compensation point, LCP, nor the light saturating value, Is varied significantly
 June 2005 and 2006July 2005 and 2006August 2005 and 2006
Amax (μmol m−2 s−1)13.2 ± 1.2a19.7 ± 1.2b,c17.1 ± 1.1a,b22.8 ± 1.3c14.2 ± 1.2a19.2 ± 1.3b,c
Amax (nmol g−1 s−1)153.1 ± 17.6a247.8 ± 15.6b,c188.8 ± 11.8a,b286.8 ± 16.7c156.9 ± 13.9a242.3 ± 16.4b,c
Aqe0.0027 ± 0.00060.0021 ± 0.00020.0021 ± 0.00020.0017 ± 0.00020.0024 ± .00040.0018 ± 0.0002
Rd (μmol m−2 s−1)−2.33 ± 0.86−1.94 ± 0.57−2.24 ± 0.34−1.99 ± 0.50−3.03 ± 0.64−2.67 ± 0.5
LCP (μmol m−2 s−1)68 ± 1941 ± 1252 ± 744 ± 1074 ± 1268 ± 10
Is (μmol m−2 s−1)589550541782464697
Chlorophylla + b (mg g−1dry wt)0.98 ± 0.06a2.81 ± 0.27b1.55 ± 0.09a3.76 ± 0.22c1.35 ± 0.09a3.20 ± 0.24b,c

Amax was the only parameter from LRC to be significantly different between Phragmites lineages (F5, 61 = 7.38, P < 0.0001) (2 × 3 factorial anova). Throughout the growing season, introduced Phragmites consistently maintained greater rates of light-saturated photosynthesis (33–42%), with the introduced type having greater photosynthetic rates (plant effect) (F1,61 = 14.02, P < 0.0001), which varied throughout the growing season (month effect) (F2,61 = 3.17, P = 0.049), and a significant interaction of plant type and time of year (F5,61 = 6.67, P < 0.0001) (Table 1). Rates of photosynthesis mirror the patterns in foliar chlorophyll concentrations (Table 2). Post hoc contrasts indicate that introduced Phragmites had significantly greater photosynthetic rates than the native type within each month. When Amax was normalized to a mass basis, introduced Phragmites sustained rates of photosynthesis that were more than 51% greater than that of the native type (Table 2). To sustain these higher rates of photosynthesis, introduced Phragmites had greater light requirements (Table 2), nearly double the rates of stomatal conductance (Fig. 3) and as a consequence, lower water use efficiency (WUE) (assimilation:transpiration) 3.48 ± 0.12 μmol (CO2) mmol−1 (H2O) vs. 4.02 ± 0.18 μmol (CO2) mmol−1 (H2O) (F1,55 = 6.16, P = 0.016). No differences were observed in LCP (F5, 61 = 0.74, P = 0.59) or Aqe (F5, 61 = 0.97, P = 0.45) (Table 2).

Figure 3.

 Plot of photosynthesis vs. stomatal conductance on native (a) and introduced (b) Phragmites during light saturated conditions (c. 1500 PPFD) during the 2005 and 2006 field seasons. Plotted line indicates linear regression of best fit.

Similar patterns were observed between the lineages in glasshouse-grown plants, with the introduced Phragmites exhibiting 80% greater values of Amax (F1,23 = 20.37, P = 0.0002) and double the concentration of photosynthetic pigments (F1,163 = 93.96, P < 0.0001) (Table 3). There was no significant effect of year on the analysed parameters (F1,23 = 2.62, P = 0.12). While there were seasonal differences observed in dark plant respiration (Rd), no significant differences were observed between the genetic lineages in both field- and glasshouse-grown plants (Tables 2 and 3).

Table 3.   Photosynthetic parameter estimates and chlorophyll concentrations for glasshouse-grown native and introduced Phragmites plants (= 24 total light response curves and = 168 for pigment analysis). Only Amax (P = 0.002) and chlorophyll concentrations (P < 0.0001) varied significantly. Neither quantum efficiency, Aqe, dark respiration, Rd, the light compensation point, LCP nor the light saturating value, Is varied significantly
 Glasshouse 2004 and 2005
Amax (μmol m−2 s−1)9.6 ± 0.817.3 ± 1.5
Aqe0.0039 ±  .00080.0034 ± 0.0004
Rd (μmol m−2 s−1)−1.17 ± 0.82−1.5 ± 1.12
LCP (μmol m−2 s−1)9 ± 66 ± 9
Is (μmol m−2 s−1)230293
Chlorophylla + b (mg g−1dry wt)2.76 ± 0.145.56 ± 0.2


Previously described morphological differences between North American and introduced Phragmites lineages (Blossey 2002) can largely be attributed to differences in foliar SLA and photosynthetic pigment concentrations. The greater SLA and greater foliar chlorophyll content of the introduced lineage contribute to its previously described phenotypic differences (e.g. darker leaf colour) from the native lineage. We believe these ecophysiological differences to be genetically inherited due to the consistent phenotypic colour differences observed in the field across a wide range of environmental conditions (Blossey 2002), and to consistency in results between in situ and glasshouse studies. Differences in SLA may also help explain the previously described differences in plant biomass and relative growth rates (RGR) between the two lineages (Vasquez et al. 2005). Our results are contrary to differences in Rubus spp. where non-invasive genotypes had a greater SLA (McDowell 2002). Plants with lower SLA have a greater carbon construction cost per unit LA (Poorter & De Jong 1999). It has been reported that SLA, not photosynthesis, was the predominant factor in variations of RGR among cultivars of tree species (Ceulemans 1989). As SLA did not change significantly throughout the plant canopy, this may suggest that Phragmites plants do not have shade-adapted leaves and may be able to sustain higher rates of canopy photosynthesis assuming adequate light penetration into the canopy. LRC measurements within the Phragmites canopy indicated that light-saturated photosynthetic rates are constant (Hager & Knops 2003), which may be attributed to their potential for carbon gain and invasive potential.

Ecophysiologically, the introduced lineage is superior to the native type under current environmental conditions. We present ecophysiological differences between conspecific genetic lineages, contrary to those reported in the invasive species, P. arundinacea, where ecophysiological differences could not be attributed to the invasive potential (Brodersen, Lavergne & Molofsky 2008). Although increased ploidy levels may result in increased photosynthetic rates and differences in leaf morphology (Hull-Sanders et al. 2009), native and introduced Phragmites are likely the same ploidy level (Saltonstall et al. 2007). Throughout the growing season, introduced Phragmites consistently maintains an at least 51% higher photosynthetic rate per unit mass than native Phragmites. These higher rates may be attributed to greater chlorophyll per unit LA. Photosynthetic differences between lineages are consistent with the differences observed in chlorophyll concentration and SLA. The lower SLA and the lower chlorophyll content in the native type may explain the lower photosynthetic rates due to thicker leaves and lower light penetration to the chloroplasts. Plants with lower SLA are associated with greater WUE (Van den Boogaard & Villar 1998), which is the case for native Phragmites. The increased WUE is corroborated by the 50–100% lower rates of stomatal conductance in the native lineage, and the greater slope in photosynthesis : stomatal conductance (Fig. 3). By contrast, for the introduced type to maintain such a high rate of photosynthesis, it must sacrifice water loss for increased C gain, resulting in lower instantaneous WUE. Greater values of stomatal conductance are also associated with greater photosynthetic rates (Wong, Cowan & Farquhar 1979). Differences in the size and densities of stomata between native and introduced lineages (Saltonstall et al. 2007) may complicate the interpretation of these data. Differences in stomatal conductance as related to differences in stomatal aperture and density, merit future investigations.

Corroborative field and glasshouse results indicate that Atlantic coast native and introduced Phragmites lineages are physiologically different and that the differences observed cannot be attributed to microsite differences. While clone-specific differences have been reported in Phragmites (Rolletschek et al. 1999a,b; Lessmann et al. 2001; Hansen et al. 2007), it is not known to which genetic lineage (sensuSaltonstall 2002) the clones belong, making interpretation of previous studies impossible. Although Hansen et al. (2007) reported no differences in ecophysiological parameters among North American Atlantic Coast and introduced clones, both clones were indeed the introduced genetic lineage (H. Brix, pers. comm.). Morphological differences in Phragmites are highly conserved (Lessmann et al. 2001), and differences in SLA and chlorophyll content support our hypothesis of genetic differences between lineages.

The combination of greater Amax rates, greater photosynthetic canopy, greater SLA and an extended growing season increase the invasion potential of introduced Phragmites. The 51% greater Amax rates are compounded by the 38–83% greater photosynthetic canopy in the introduced lineage. The greater foliar canopy in the introduced lineage is consistent with the results of Lynch et al. 2006 and is similar to that of other invasive plant species (Feng, Wang & Sang 2007). Additionally, due to a greater SLA and a thus lower carbon construction cost, more photosynthate may also be stored below ground increasing the potential for clonal expansion.

Phenological differences may further amplify the physiological advantages (greater Amax, canopy, SLA) of introduced Phragmites. While both lineages begin growing at the same time of year, patterns of emergence are quite different. League et al. (2006) reported that introduced Phragmites emerges at 13-fold greater stem densities in Delaware wetlands in March. More importantly, the introduced lineage photosynthesizes at nearly the same rate throughout the year, compared to the native type which varies by 30% over the course of the growing season (Table 2). In addition to this vigorous start, the introduced lineage also has an extended growing season and does not begin senescing until September or October, whereas Atlantic Coast native lineages begin senescing in August (Blossey 2002; T. J. Mozdzer, Pers. observ.). This extended phenology (up to 2 months) and nearly constant leaf-level photosynthetic rates afford the introduced type additional time to increase above-ground biomass and below-ground reserves. An extended phenology in invasive Lythrum salicaria genotypes was attributed to increased invasive potential in subsequent years (Chun et al. 2007). Similar to invasive congeners (McDowell 2002), introduced Phragmites continues to thrive weeks to months after native Phragmites and other native plants senesce (T. J. Mozdzer, pers. observ.), resulting in a greater carbon gain and the potential to increase its below-ground reserves. With such dramatic differences in photosynthetic potential and phenology, the question remains: why has the invasion of introduced Phragmites taken so long, if this non-native type has been in North America for over a century?

We propose that the introduced lineage was not a good ecological competitor until environmental conditions became favourable for its success. While several studies have correlated anthropogenic N eutrophication with Phragmites invasion (Bertness, Ewanchuk & Silliman 2002; King et al. 2007), and it has been demonstrated that Phragmites demands 50% more N than the species it is replacing in freshwater, brackish, and salt marshes (Templer, Findlay & Wigand 1998; Windham & Ehrenfeld 2003; Windham & Meyerson 2003), no studies have compared N demand between the Phragmites lineages. A simple N demand calculation can be made from our data by multiplying the mean above-ground biomass of native and introduced Phragmites and the mean foliar N content of native and introduced lineages (2.3% and 2.8%, respectively) (from this study, Saltonstall & Stevenson 2007; Packett & Chambers 2006), yielding N demands of 14.5 and 51.7 g plant N m−2 year−1, respectively for the native and introduced lineages. Our data suggest that at ambient plant densities, introduced Phragmites demands nearly four times more N to support its above-ground biomass than the native type. Under current conditions of higher N availability (Galloway et al. 2004), the introduced lineage can exploit increased N pools due to greater N uptake rates (Mozdzer, Zieman & McGlathery in press), fuelling the photosynthetic machinery, which results in the currently observed higher rates of photosynthesis and RGR (Vasquez et al. 2005). On the other hand, our data also suggest that the native type may be a low-nutrient specialist, supported by its lower RGR (Vasquez et al. 2005), lower N demand, lower SLA and greater WUE. The combination of global change and physiological differences may be a factor facilitating the shift in dominant Phragmites lineage in North American Atlantic coast wetlands.

We propose that genetically inherited ecophysiological and morphological differences between lineages explain the success of the introduced type under current environmental conditions, and that the native lineage cannot exploit eutrophied habitats as well as the introduced type. This is the first study to demonstrate how conspecific physiological and morphological differences can be attributed to the biological invasion of an introduced lineage. We suggest that global changes, largely changes in N availability, are facilitating the current invasion due to the introduced type’s high N demand and the native type’s physiological inability to exploit this resource. What remains unknown is how global change, namely rising CO2 concentration, affects the ecophysiology of this species and the individual lineages. Investigations are currently under way to systematically test the effects of global change on this species, which can be used as a potential model invader due to the plants’ cosmopolitan distribution and distinct genetic lineages.


We would like to thank the NSF grant DEB 0621014 for funding, J. W. Snyder, M. Takahashi and B.C. Curtis for field and laboratory assistance, as well as T.M. Smith, M. Lerdau, J.A. Langley and two anonymous referees for their critical review of an earlier version of the manuscript.