Differences in plasticity between invasive and native plants from a low resource environment


*Correspondence author. Department of Biological Sciences, Chapman University, Orange, CA 92866. E-mail: jlfunk@chapman.edu


  • 1Phenotypic plasticity is often cited as an important mechanism of plant invasion. However, few studies have evaluated the plasticity of a diverse set of traits among invasive and native species, particularly in low resource habitats, and none have examined the functional significance of these traits.
  • 2I explored trait plasticity in response to variation in light and nutrient availability in five phylogenetically related pairs of native and invasive species occurring in a nutrient-poor habitat. In addition to the magnitude of trait plasticity, I assessed the correlation between 16 leaf- and plant-level traits and plant performance, as measured by total plant biomass. Because plasticity for morphological and physiological traits is thought to be limited in low resource environments (where native species usually display traits associated with resource conservation), I predicted that native and invasive species would display similar, low levels of trait plasticity.
  • 3Across treatments, invasive and native species within pairs differed with respect to many of the traits measured; however, invasive species as a group did not show consistent patterns in the direction of trait values. Relative to native species, invasive species displayed high plasticity in traits pertaining to biomass partitioning and leaf-level nitrogen and light use, but only in response to nutrient availability. Invasive and native species showed similar levels of resource-use efficiency and there was no relationship between species plasticity and resource-use efficiency across species.
  • 4Traits associated with carbon fixation were strongly correlated with performance in invasive species while only a single resource conservation trait was strongly correlated with performance in multiple native species. Several highly plastic traits were not strongly correlated with performance which underscores the difficulty in assessing the functional significance of resource conservation traits over short timescales and calls into question the relevance of simple, quantitative assessments of trait plasticity.
  • 5Synthesis. My data support the idea that invasive species display high trait plasticity. The degree of plasticity observed here for species occurring in low resource systems corresponds with values observed in high resource systems, which contradicts the general paradigm that trait plasticity is constrained in low resource systems. Several traits were positively correlated with plant performance suggesting that trait plasticity will influence plant fitness.


Understanding the factors that contribute to the success of invasive species may facilitate the prediction of future invasions, determine the best ways to control invasive species, and elucidate the impact of invasive species on native communities (Pysek & Richardson 2007). To identify traits associated with invasiveness, researchers have primarily evaluated invasive and native or introduced, non-invasive species within large regional data sets (Sutherland 2004; Hamilton et al. 2005) and within plant genera or families (Grotkopp et al. 2002; Muth & Pigliucci 2006). A recent analysis of these two approaches concluded that plant height, vegetative growth, and early and extended flowering are strongly associated with invasiveness (Pysek & Richardson 2007). However, our understanding of how the dynamic nature of reproductive, life-history, physiological and morphological traits contributes to invasiveness is still limited.

It has been repeatedly suggested that invasive species possess high phenotypic plasticity (broadly defined as the ability of organisms to alter their morphology and/or physiology in response to varying environmental conditions) which may allow them to occupy a wide range of new environments (Baker 1965; Marshall & Jain 1968; Sultan 2001; Callaway et al. 2003; Daehler 2003; Pigliucci 2005; Rejmanek et al. 2005). While empirical studies generally support this idea (e.g. Rice & Mack 1991; Pattison et al. 1998; Gerlach Jr. & Rice 2003; Niinemets et al. 2003), the lack of phylogenetically and ecologically equivalent comparisons among invasive and native species limits the ability to explicitly link plasticity with invasiveness. Phylogenetic comparative designs are necessary to minimize trait differences associated with comparing unrelated species and disparate life forms (Burns & Winn 2006; Muth & Pigliucci 2006; Richards et al. 2006).

Because plasticity in physiological and morphological traits can provide greater access to limiting resources, invasive species may benefit from plasticity in low resource environments (where plant growth is strongly limited by water, nutrient or light availability). For example, invaders that adjust physiological or morphological traits to take advantage of spatially variable light availability (e.g. sunflecks) or temporally variable water and nutrient availability (e.g. storms, deposition) in low resource systems may outperform less plastic neighbouring species (Poorter & Lambers 1986; Davis et al. 2000). However, Chapin (1980) and Grime et al. (1986) argued that resource conservation (e.g. storage, retention of existing leaves) should be more advantageous than resource acquisition (e.g. the deployment of new leaves or roots) in low resource or stressful environments. Thus, native species adapted to low resource systems may exhibit low phenotypic plasticity. While recent work supports this idea (e.g. Balaguer et al. 2001; Steinger et al. 2003), it is not known if species invading low resource systems also demonstrate low trait plasticity or if they succeed in these systems due to increased trait plasticity.

I examined how light and nutrient availability influenced physiological and morphological plasticity among five phylogenetically related pairs of invasive and native species that occur in nitrogen-limited habitats in Hawai’i. Based on the theory of Chapin (1980) and Grime et al. (1986), I predicted that invaders would be similar to natives adapted to these low resource systems by displaying low trait plasticity. A recent field survey of species across low resource habitats in Hawai’i found that invasive species displayed similar or high resource-use efficiency (resource conservation traits) relative to neighbouring native species (Funk & Vitousek 2007). Thus, I also predicted that invasive species would display high resource-use efficiency and that, across species, resource-use efficiency and plasticity would be negatively correlated.

As trait plasticity will only influence the success of invasive species if traits are linked to plant fitness, I also explored the functional significance of the observed plasticity. This study examined species-level plasticity (plasticity among individuals within a population across different environments; Richards et al. 2006; Valladares et al. 2006) as opposed to genotype-level plasticity (the capacity of individual genotypes to produce different phenotypes across different environments; Sultan 2000). While both species- and genotype-level approaches are useful in the exploration of invasive species success (Richards et al. 2006), species-level approaches permit the measurement of more traits and species by sacrificing replication at the level of individual genotype. In the absence of genetic information required for the determination of adaptive plasticity by traditional selection analysis (e.g. Dudley 1996; Heschel et al. 2004), I evaluated the functional significance of species-level plasticity by correlating traits with plant performance across different environments (Donovan & Ehleringer 1994; Burton & Bazzaz 1995; Funk et al. 2007).


experimental design

From January to May 2005, individuals from five phylogenetically related pairs of native and introduced invasive species were propagated from seed or cuttings (Table 1) and grown in the University of Hawai’i, Volcano Agricultural Experiment Station glasshouse (19°28′ N, 155°16′ W, 1240 m). The glasshouse was open on two sides permitting wind and insects to enter. Light and temperature were not controlled and approximated ambient conditions. The selected species are co-occurring common native and invasive species within Hawai’i Volcanoes National Park, located on the island of Hawai’i (Table 1) and were selected to represent a broad taxonomic diversity. These species occur on young volcanic soils that are nitrogen-poor (Vitousek et al. 1993). Species within two pairs (Myrtaceae and Rosaceae) also occur in light-limited rainforest. All five introduced species are considered invasive weeds (A Global Compendium of Weeds, <http://www.hear.org/gcw/index.html>).

Table 1.  Species names, propagule type, initial plant size and population source data for the 10 species used in this study. All seeds and cuttings were collected from multiple (> 24) individuals within single, large populations. For cuttings, the mean initial plant size ± 1 SE across the four treatments are given. An asterisk (*) indicates invasive species
FamilySpecies typeScientific namePropagule typeInitial plant size (cm)Population source (elevation)
  1. Abbreviations: Hawai’i Volcanoes National Park (HAVO).

FabaceaeN2-fixing treeLeucaena leucocephala*seedKailua-Kona, HI (70 m)
 Acacia koaseedMauna Loa, HAVO (1300–1800 m)
MyrtaceaeTreePsidium cattleianum*cutting7.85 ± 0.08Kilauea, HAVO (1200 m)
 Metrosideros polymorphacutting9.72 ± 0.15Kilauea, HAVO (1200 m)
PoaceaeC3 grassHolcus lanatus*seedMauna Loa, HAVO (2070 m)
 Deschampsia nubigenaseedMauna Loa, HAVO (2070 m)
PoaceaeC4 grassPaspalum urvillei*seedMauna Loa, HAVO (1370 m)
 Eragrostis variabilisseedMauna Loa, HAVO (1370 m)
RosaceaeShrubRubus ellipticus*cutting17.81 ± 0.48Volcano, HI (1200 m)
 Rubus hawaiiensiscutting13.75 ± 0.31Volcano, HI (1200 m)

Twenty-four plants per species were first sowed into 0.5 L pots with a 3 : 2 mixture of cinder and potting mix (Sunshine Retail mix #1, SunGro Horticulture, Seba Beach, AB, Canada), which closely resembles the nitrogen-poor soil in the area (S. McDaniel, personal communication). In July, plants were transferred to 2 L pots with the same soil mixture. A systemic insecticide was added to all plants at the beginning of the experiment (Marathon 1% granular, Olympic Horticultural Products, Mainland, PA). No herbivory was observed during the 6-month study.

Plants were subjected to four growth environments (low light/low nutrients, low light/high nutrients, high light/low nutrients, and high light/high nutrients). There were six individuals per species per treatment. The selected light environments represent open mixed Acacia/Metrosideros forest (400–600 µmol m−2 s−1; Denslow et al. 2006) and open, unforested areas (1800–2000 µmol m−2 s−1; roadsides, forest clearings) typically encountered by these species. Light quality (red : far-red ratios, R : FR) was altered to mirror conditions in nearby montane forests (R : FR 0.62 for low light treatment, R : FR 0.89 for high light treatment). To create the low light environment, sheets of plastic were strung across the top and sides of a wooden frame 1.5 m above the bench in the glasshouse. Within a light treatment, plants were positioned randomly on the bench and were rotated every month. Ambient temperature did not vary across light treatments. Plants were fertilized with a 24–8–16 (N–P–K plus micronutrients, percent by mass) fertilizer (Miracle Gro, The Scotts Company, Marysville, OH). Low nutrient plants received monthly doses of 0.035 g N for a total of 0.175 g N. High nutrient plants received 10 applications of fertilizer over 5 months: 0.035 g N (first 5 doses) and 0.070 g N (last 5 doses) for a total of 0.525 g N. There was no plant mortality over the 6-month duration of the experiment.

A suite of leaf-level (physiological, morphological) and plant-level (biomass allocation) traits were measured. Leaf-level traits were selected based on their expected functional significance in low resource habitats (e.g. resource use efficiency). I also selected traits that are known to respond to variation in light and nutrient availability (e.g. biochemical measures) as well as a measure of light stress (photoinhibition). As reproductive measures are difficult to obtain during short-term studies of perennial species, total above- and below-ground biomass was used to evaluate plant performance (Farris & Lechowicz 1990; Sultan et al. 1998; Givnish 2002). See Table 2 for abbreviations and descriptions of traits.

Table 2.  Abbreviations and descriptions of the traits measured in the study
Plant-level traits
 LWRLeaf weight ratiog leaf g−1 plant
 SWRStem weight ratiog stem g−1 plant
 RWRRoot weight ratiog root g−1 plant
 LARLeaf area ratiom2 leaf kg−1 plant
 R/SRoot to shoot ratiog root g−1 stem and leaf
Leaf-level traits
 AsatPhotosynthetic rate at saturating lightµmol COm−2 leaf s−1
 WUEWater use efficiencyµmol COmol−1 H2O
 PNUEPhotosynthetic nitrogen use efficiencyµmol COmol−1 N s−1
 NareaLeaf nitrogen content on area basisg N m−2 leaf
 NmassLeaf nitrogen content on mass basismg N g−1 leaf
 LMALeaf mass per areag leaf m−2 leaf
 Chl a/bRatio of chlorophyll a/b
 Chl/NRatio of chlorophyll to nitrogen content
 VmaxMaximum rate of carboxylationµmol COm−2 leaf s−1
 ΦPSIIEffective quantum yield of PSIIΔF/inline image
 Fv/FmRatio of variable to maximum fluorescence
Performance measure
 BiomassTotal above- and below-ground biomassg plant

leaf-level measures

Physiological traits were monitored in leaves produced after the initiation of treatments to eliminate any confounding effects of species differences in physiological plasticity in existing leaves (e.g. Newell et al. 1993). All leaves were fully expanded, recently mature leaves. In September, photosynthetic rates, transpiration rates and chlorophyll fluorescence were measured with a LI-6400 portable photosynthesis system with a fluorescence chamber (LI-COR, Lincoln, NE). All measures were conducted at two light levels, 600 and 1900 µmol photon m−2 s−1, at 400 µL L−1 CO2 and under ambient conditions of humidity and temperature. Water use efficiency (WUE) was calculated as the ratio of photosynthesis to transpiration at 1900 µmol photon m−2 s−1.

The effective quantum yield of PSII (ΦPSII), measured at 600 µmol photon m−2 s−1, was calculated as ((inline image − Fs)/inline image), where Fs is the fluorescence yield of a light-adapted leaf and inline image is the maximal fluorescence during a saturating light flash. Pre-dawn fluorescence measurements were taken to assess the fraction of absorbed photons that are used for photochemistry (Fv/Fm) which was calculated as (Fm − Fo/Fm), where Fo is the fluorescence in total darkness and Fm is fluorescence of a dark-adapted leaf during a saturating light flash. Fv/Fm is around 0.8 for healthy plants and can be used as an indicator of photoinhibition (Cavender-Bares & Bazzaz 2004). CO2 response curves were determined by varying chamber CO2 concentration between 0 and 900 µL L−1 while the light level was 1900 µmol photon m−2 s−1. Maximum carboxylation rates (Vmax) were estimated from the CO2 response curves when intercellular CO2 concentration was below 200 µL L−1 following Wullschleger (1993) and based on the photosynthesis model of Farquhar et al. (1980). Vmax was not measured on C4 grasses.

Three to four leaf disks (0.54 to 0.65 cm2) were collected from each plant, immediately frozen, and stored at −80 °C. Leaves were ground in 100% acetone in a chilled mortar with a small amount of quartz sand and MgCO3 to prevent acidification. Samples were centrifuged for 2 min at 3000 rpm and the absorbance of the supernatant was measured with a UV/VIS spectrophotometer (Beckman DU-640, Beckman Coulter, Inc., Fullerton, CA). Chlorophyll a and b were determined using a multi-wavelength analysis at 470, 645, 662, and 710 nm (Lichtenthaler & Buschmann 2001). High Chl/N and low Chl a/b ratios indicate increased allocation to light harvesting components (Evans 1988).

After photosynthetic measures, leaves were collected, scanned to determine leaf area, dried at 65 °C for 3 days, and weighed to determine leaf mass per area (LMA). Ground leaves were analyzed for leaf N content with an elemental analyzer (CE Instruments Flash EA 1112, CE Elantech, Lakewood, NJ, USA). Photosynthetic nitrogen use efficiency (PNUE) was calculated as the ratio of photosynthesis to leaf N.

plant-level measures

After 6 months of growth, all plants were harvested. Above-ground biomass of trees and shrubs was separated into stems and leaves. Leaves were scanned and total leaf area for each plant was determined (senescent leaves were excluded). Roots were carefully separated from soil and washed to minimize fine root loss. All material was dried and weighed to determine leaf weight ratio (LWR), stem weight ratio (SWR), root weight ratio (RWR), leaf area ratio (LAR), root to shoot ratio (R/S), and total biomass. Because grass species lacked a clear distinction between shoots and leaves (i.e. much of the shoot was elongated leaf sheaths), grass shoots were treated as leaves and were incorporated into LWR. SWR was not measured for grasses.

plasticity indices and statistical analysis

To examine differences in leaf-level traits, I used a mixed, nested anova with origin (introduced or native), light and nutrient as fixed factors and species (nested within origin) as a random factor. The F-ratio for origin was evaluated using species mean square in the denominator. Because plant size can influence above- and below-ground patterns of biomass partitioning (e.g. Coleman & McConnaughay 1995), a mixed, nested ancova (with the same main effects and biomass as a covariate) was used to evaluate treatment differences in plant-level traits. Following Moran (2003), sequential Bonferroni corrections for multiple statistical tests were not conducted. As recommended by Moran (2003), all P-values are reported. Paired t-tests were used to compare trait values within species pairs.

To compare the degree of plasticity among species and traits, I generated a plasticity index (PIV) for each trait (Valladares et al. 2006). The index ranges from zero (no plasticity) to one (maximum plasticity) and is the difference between the minimum and maximum value of the treatment means of a trait divided by the maximum value. Differences in PIV among native and invasive species for a given trait were evaluated with interaction terms (origin × light, origin × nutrient, origin × light × nutrient) in the nested anova/ancova. A one-way anova was used to assess differences in PIV among leaf- and plant-level traits.

Pearson product-moment correlation coefficients were generated to evaluate the linear association between traits and biomass across environments and between resource conservation traits (PNUE, WUE, ΦPSII, LMA) and species mean PIV. All analyses were performed in JMP 5.1.2 (SAS Institute Inc., Cary, NC). Data that violated the anova assumptions of normality and homogeneity of variance were Box Cox transformed.


species effects

Within pairs, several invasive and native species differed with respect to the 16 traits measured (Table 3) but invasive species as a group did not show consistent differences in the direction of trait values relative to grouped natives (Table 4). Three invasive species (L. leucocephala, H. lanatus, P. urvillei) allocated more biomass to roots (higher RWR, R : S; lower LWR) and had thinner leaves (LMA) than did paired natives (Table 3). Two natives (M. polymorpha, A. koa) and one invader (R. ellipticus) had higher photosynthetic rates than their paired species. Nitrogen use patterns varied across species with two invasives (R. ellipticus, P. cattleianum) and three natives (A. koa, D. nubigena, E. variabilis) displaying higher Narea and two invasives (H. lanatus, P. urvillei) and two natives (M. polymorpha, A. koa) displaying higher PNUE than their paired species (Table 3).

Table 3.  Leaf- and plant level traits for five pairs of phylogenetically related invasive and native species across light and nutrient treatments. Values are means with one standard error. Sample sizes are n = 5 for leaf-level traits and n = 6 for plant-level traits. Significantly higher values (P < 0.05) within a native/invasive species pair as measured by paired t-tests are denoted by bold font. Trait abbreviations are given in Table 2. An asterisk (*) indicates invasive species. (–) Denotes no data collected
 TreeN2-fixing treeShrubC3 grassC4 grass
Metrosideros polymorphaPsidium* cattleianumAcacia koaLeucaena* leucocephalaRubus hawaiiensisRubus* ellipticusDeschampsia nubigenaHolcus* lanatusEragrostis variabilisPaspalum* urvillei
Plant-level traits
 LWR0.58 (0.01)0.50 (0.01)0.35 (0.01)0.21 (0.01)0.38 (0.02)0.39 (0.01)0.73 (0.03)0.63 (0.03)0.85 (0.02)0.70 (0.02)
 SWR0.22 (0.01)0.27 (0.01)0.32 (0.01)0.28 (0.01)0.26 (0.01)0.25 (0.02)
 RWR0.20 (0.01)0.23 (0.01)0.28 (0.01)0.51 (0.02)0.37 (0.02)0.36 (0.02)0.27 (0.03)0.37 (0.03)0.15 (0.02)0.30 (0.02)
 LAR4.59 (0.25)3.98 (0.19)6.92 (0.54)5.46 (0.62)8.37 (0.71)5.67 (0.44)6.52 (0.65)17.02 (1.55)16.66 (1.37)16.04 (1.18)
 R/S0.25 (0.02)0.30 (0.03)0.37 (0.02)1.13 (0.10)0.62 (0.06)0.62 (0.07)0.40 (0.06)0.63 (0.07)0.18 (0.02)0.47 (0.06)
Leaf-level traits
 Asat11.19 (0.72)9.07 (0.61)13.93 (0.66)7.28 (1.19)5.29 (0.49)8.90 (0.78)10.41 (1.22)9.82 (0.85)19.84 (1.47)19.93 (1.43)
 WUE4.89 (0.23)5.25 (0.36)4.30 (0.27)5.22 (0.64)3.66 (0.32)4.76 (0.19)3.23 (0.45)3.73 (0.32)8.21 (0.38)8.60 (0.23)
 PNUE123.7 (8.8)84.55 (9.06)148.9 (12.6)100.3 (11.6)102.6 (10.6)120.6 (7.4)114.4 (12.1)268.6 (28.4)324.3 (22.7)472.6 (19.6)
 Narea1.30 (0.06)1.70 (0.11)1.43 (0.10)0.94 (0.08)0.78 (0.05)1.03 (0.05)1.39 (0.12)0.57 (0.06)0.88 (0.06)0.59 (0.04)
 Nmass0.99 (0.04)1.31 (0.07)2.58 (0.09)2.29 (0.25)1.68 (0.12)1.43 (0.09)1.15 (0.08)1.52 (0.19)1.71 (0.17)1.35 (0.12)
 LMA135.7 (8.3)132.6 (8.0)56.30 (3.74)43.70 (2.97)48.47 (2.78)77.02 (4.98)125.7 (11.2)40.11 (2.63)57.29 (4.27)46.20 (2.13)
 Chl a/b2.77 (0.03)2.82 (0.03)2.93 (0.04)2.75 (0.05)2.62 (0.02)2.89 (0.04)2.68 (0.02)2.79 (0.03)3.36 (0.07)3.39 (0.02)
 Chl/N3.63 (0.14)3.56 (0.20)4.40 (0.56)4.01 (0.40)3.01 (0.28)2.76 (0.20)2.20 (0.15)4.69 (0.42)4.64 (0.43)4.71 (0.28)
 Vmax57.58 (4.14)54.25 (2.44)52.80 (4.37)46.18 (8.16)23.80 (2.99)46.21 (4.24)61.38 (6.11)39.59 (4.75)
 ΦPSII 6000.39 (0.01)0.30 (0.01)0.36 (0.02)0.27 (0.03)0.19 (0.01)0.28 (0.02)0.31 (0.02)0.26 (0.02)0.33 (0.02)0.30 (0.02)
 Fv/Fm0.81 (0.01)0.80 (0.01)0.79 (0.01)0.78 (0.01)0.80 (0.00)0.81 (0.01)0.80 (0.00)0.80 (0.00)0.75 (0.01)0.77 (0.01)
Performance measure
 Biomass7.71 (1.75)8.87 (1.16)15.04 (2.18)8.93 (2.02)17.01 (2.80)25.55 (2.79)16.51 (4.10)22.24 (3.11)12.79 (2.51)23.89 (2.96)
Table 4.  Mixed, nested (A) ancova and (B) anova results for traits and performance measured in five pairs of phylogenetically related invasive and native species across light and nutrient treatments. For plant-level traits, biomass was used as a covariate to isolate allometric effects on biomass partitioning. Numerator and denominator degrees of freedom are given for each effect (F-ratio subscript). See Table 2 for trait abbreviations. Statistically significant values (P < 0.01) are denoted by bold font
 OriginLightNutrientLight × NutrientOrigin × LightOrigin × NutrientOrigin × Light × Nutrient
A. ancova
LWR0.461,80.51520.371,172< 0.00141.461,172< 0.001 0.381,1720.537 3.171,1720.0777.011,1720.0090.001,1720.990
SWR0.141,40.72485.381,107< 0.0010.271,1070.604 0.301,1070.588 0.761,1070.3874.181,1070.0430.771,1070.382
RWR2.931,80.12587.411,174< 0.00137.131,174< 0.001 0.011,1740.907 3.451,1740.06511.581,174< 0.0010.181,1740.673
LAR0.001,80.948271.21,174< 0.00120.031,174< 0.001 0.751,1740.388 2.761,1740.0986.241,1740.0130.021,1740.885
R/S2.991,80.12283.151,174< 0.00144.241,174< 0.001 0.001,1740.967 1.401,1740.2388.941,1740.0030.221,1740.643
B. anova
Asat0.111,80.7460.611,1830.43659.671,183< 0.001 3.241,1830.074 0.731,1830.3961.531,1830.2180.021,1830.875
WUE0.331,80.5830.571,1830.45029.851,183< 0.001 0.081,1830.777 1.791,1830.1830.821,1830.3700.111,1830.738
PNUE0.161,80.7003.851,1740.0520.371,1740.545 2.731,1740.100 1.671,1740.1980.271,1740.6031.171,1740.281
Narea0.921,80.36510.591,1750.001142.91,175< 0.001 0.061,1750.809 2.021,1750.15817.231,175< 0.0012.771,1750.098
Nmass0.011,80.929154.31,175< 0.001135.51,175< 0.001 4.201,1750.042 0.821,1750.36817.581,175< 0.0012.911,1750.090
LMA0.501,80.498351.81,175< 0.0013.371,1750.068 6.021,1750.015 0.271,1750.6030.271,1750.6060.171,1750.684
Chl a/b0.161,80.703118.41,143< 0.0012.891,1430.092 0.731,1430.394 1.331,1430.2500.201,1430.6531.791,1430.183
Chl/N0.651,80.445101.91,134< 0.0011.571,1340.212 4.131,1340.044 0.011,1340.9350.591,1340.4420.921,1340.340
Vmax0.071,60.8062.571,820.11334.151,82< 0.001 0.001,820.961 1.371,820.2459.441,820.0031.751,820.189
ΦPSII0.831,80.3892.541,1830.113136.11,183< 0.001 4.511,1830.035 0.111,1830.74112.951,183< 0.0010.561,1830.454
Fv/Fm0.021,80.897245.61,182< 0.00145.221,182< 0.001 8.661,1820.004 0.661,1820.41612.681,182< 0.0011.661,1820.200
Biomass0.651,80.44363.471,175< 0.001364.51,175< 0.00118.541,175< 0.00116.111,175< 0.0015.371,1750.0224.041,1750.046


Performance patterns across the four environments are presented in Fig. 1. High light and nutrient availability increased total biomass, with nutrient availability having a stronger effect than light availability (Tables 4 and 5). Across the five species pairs, invasive species responded more favorably to nutrient availability (origin × nutrient interaction, Table 4) while native species responded more favorably to light availability (origin × light interaction, Table 4).

Figure 1.

Performance (above- and below-ground biomass) for (a) pooled native and invasive species and (b–f) phylogenetically related pairs of native (closed circles) and invasive (open circles) species. Treatments are arranged on the x-axis with the most stressful condition (low nutrient, low light) on the left and the most favourable condition (high nutrient, high light) on the right. Data are means and standard error. Significant differences (P < 0.05) in biomass between native and invasive species within a treatment are denoted by an asterisk (*).

Table 5.  Leaf- and plant level traits for native and invasive species across four treatments differing in light (L) and nutrient (N) availability. Values are means for five native and five invasive species with one standard error. Trait values not connected by the same letter (a–g) are statistically different at P < 0.05. Trait abbreviations are given in Table 2
 Native speciesInvasive species
Lo N, Lo LLo N, Hi LHi N, Lo LHi N, Hi LLo N, Lo LLo N, Hi LHi N, Lo LHi N, Hi L
Plant-level traits
 LWR 0.57 (0.10)bcd0.53 (0.08)cd0.63 (0.12)a0.59 (0.10)ab0.47 (0.10)d0.40 (0.07)e0.57 (0.10)abc0.50 (0.09)d
 SWR0.31 (0.03)abc0.25 (0.05)de0.31 (0.04)abc0.26 (0.06)de0.28 (0.03)bd0.21 (0.02)ce0.33 (0.01)a0.24 (0.02)bcde
 RWR0.24 (0.05)ce0.33 (0.04)abd0.18 (0.04)f0.26 (0.04)ce0.36 (0.05)bc0.47 (0.04)a0.23 (0.03)def0.35 (0.06)bc
 LAR10.30 (2.65)abc6.21 (1.27)de11.77 (3.13)abc6.16 (1.41)de11.08 (3.36)bd6.26 (1.76)ce13.73 (3.95)a7.47 (2.39)ce
 R/S0.35 (0.09)ce0.51 (0.09)abd0.24 (0.06)f0.37 (0.09)ce0.63 (0.16)bc0.94 (0.16)a0.32 (0.06)cef0.62 (0.18)bc
Leaf-level traits
 Asat11.33 (2.40)abcd9.34 (1.90)bd13.84 (2.97)ac14.02 (2.95)ac8.99 (2.15)cd8.12 (1.97)cd13.04 (2.35)ab14.00 (3.59)ab
 WUE4.16 (0.91)bd4.58 (0.77)abcd5.12 (1.00)abcd5.57 (0.91)ac5.01 (0.79)abcd4.69 (0.91)cd6.16 (0.68)ab6.19 (1.06)ab
 PNUE175.0 (42.3)ab143.6 (44.4)b176.2 (38.6)a156.9 (43.1)ab229.8 (81.8)ab186.5 (67.5)ab184.8 (58.2)ab236.4 (91.9)ab
 Narea0.95 (0.11)bc1.10 (0.25)bc1.15 (0.11)bc1.42 (0.17)a0.68 (0.15)c0.76 (0.17)c1.18 (0.18)ab1.25 (0.32)ab
 Nmass1.67 (0.30)cde1.26 (0.34)fg2.09 (0.31)ab1.48 (0.27)cde1.34 (0.12)bdf1.01 (0.11)eg2.51 (0.42)ac1.46 (0.14)bdf
 LMA66.62 (15.40)cd95.82 (18.50)ab64.08 (14.91)cd112.3 (27.7)ab51.61 (12.01)bd78.45 (19.07)ac53.49 (14.49)bd88.14 (24.97)ac
 Chl a/b2.76 (0.10)bd2.95 (0.16)ac2.80 (0.11)bd2.97 (0.16)ac2.84 (0.14)cd2.99 (0.09)ab2.82 (0.13)cd3.06 (0.12)ab
 Chl/N0.42 (0.07)ac0.26 (0.02)d0.45 (0.09)ac0.28 (0.04)bd0.51 (0.05)a0.29 (0.03)cd0.45 (0.05)ab0.34 (0.05)cd
 Vmax38.98 (8.20)ab50.86 (11.89)ab50.85 (7.42)ab54.86 (9.75)ab35.00 (7.23)b32.59 (4.99)b56.88 (1.82)a61.77 (5.46)a
 ΦPSII0.28 (0.04)bcde0.28 (0.04)ce0.33 (0.03)abd0.37 (0.03)a0.22 (0.03)de0.22 (0.01)de0.34 (0.01)abc0.36 (0.02)abc
 Fv/Fm0.81 (0.01)ab0.77 (0.01)de0.81 (0.01)ab0.78 (0.01)cde0.80 (0.01)ac0.75 (0.01)e0.82 (0.01)ac0.79 (0.01)bd
Performance measure
 Biomass5.14 (0.79)e7.81 (1.49)cde12.46 (3.79)cd29.84 (3.87)ab8.77 (2.80)de8.77 (2.23)de23.54 (5.68)bc30.50 (5.22)a

In some cases (R. ellipticus, P. urvillei), invasive species outperformed natives under both the highest and the lowest resource conditions (Fig. 1). In other cases (P. cattleianum, H. lanatus), invasive species outperformed natives only under low resource conditions (Fig. 1). In one case, a native species (A. koa) outperformed the invasive under most conditions.

trait plasticity

Across all ten species, light and nutrient availability influenced many leaf- and plant-level traits (Tables 4 and 5) with some traits showing little change across treatments (e.g. Fv/Fm PIV = 0.07; Table 6) and others showing large changes (e.g. R/S PIV = 0.61; Table 6). With respect to plant-level traits, plants growing in high light environments decreased above-ground allocation (LWR, SWR, LAR) and increased below-ground allocation (RWR, R/S) while plants growing in high nutrient environments showed the opposite pattern (Tables 4 and 5). Nutrient availability had a strong effect on biomass partitioning even when allometric effects where accounted for (ancova; biomass as covariate; Table 4). With respect to leaf-level traits, high light availability resulted in thicker leaves (high LMA), decreased allocation to light harvesting components of photosynthesis (high Chl a/b, low Chl/N), high leaf N content (high Narea but low Nmass due to higher LMA), and high photoinhibition (low Fv/Fm; Tables 4 and 5). High nutrient availability resulted in high leaf N content (Nmass, Narea), photosynthetic rates, ΦPSII, WUE, carboxylation capacity (Vmax), and low photoinhibition (high Fv/Fm; Tables 4 and 5). Significant light × nutrient interactions generally corresponded with a stronger effect of nutrient addition in high light treatments.

Table 6.  Plasticity indices (PIV) for plant- and leaf-level traits measured in five pairs of phylogenetically related invasive and native species across light and nutrient treatments (n = 20 per trait per species). The index ranges from 0 to 1 (see text for PIV calculation). Species and trait PIV are mean values averaged over all species and traits, respectively. An asterisk (*) indicates invasive species. (–) Denofis no data collected
 TreeN2-fixing treeShrubC3 grassC4 grassMean PIV
Metrosideros polymorphaPsidium* cattleianumAcacia koaLeucaena* leucocephalaRubus hawaiiensisRubus* ellipticusDeschampsia nubigenaHolcus* lanatusEragrostis variabilisPaspalum* urvilleiNativeInvasive
Plant-level traits
Leaf-level traits
 Chl a/b0.
 Species PIV0.280.370.270.510.360.420.420.450.390.400.340.43

Invasive species exhibited higher plasticity for many leaf- and plant-level traits, but only in response to nutrient availability (origin × nutrient interactions, Table 4). In response to variation in nutrient availability invasive species displayed higher plasticity in above- vs. below-ground biomass partitioning (LWR, RWR, R/S), carbon assimilation (Vmax), and traits pertaining to leaf-level nitrogen (Narea, Nmass) and light use (ΦPSII, Fv/Fm; Table 4).

Across species, leaf- and plant-level traits were similarly plastic (native species: leaf-level PIv = 0.32, plant-level PIv = 0.39, P = 0.34; invasive species: leaf-level PIv = 0.41, plant-level PIv = 0.48, P = 0.46). Mean species PIV was not correlated with any measure of resource-use efficiency (PNUE, WUE, ΦPSII, LMA; all P > 0.05).

The functional significance of individual traits was assessed by correlating traits with total plant biomass across all four treatments. Several plant- and leaf-level traits correlated strongly with biomass (| r | > 0.60, Table 7). The effective quantum yield (ΦPSII) was strongly correlated (P < 0.01) with performance in two native (R. hawaiiensis, D. nubigena) and three invasive (L. leucocephala, R. ellipticus, P. urvillei) species (Table 7). Photosynthetic rate (Asat) and Narea were also strongly correlated with plant biomass in four of the five invasive species (Table 7). Besides ΦPSII, LMA was the only trait that was highly correlated with performance in more than a single native species (M. polymorpha, D. nubigena, E. variabilis).

Table 7.  Pearson product-moment correlation coefficients (r) for relationships between traits and total biomass for five pairs of phylogenetically related invasive and native species across light and nutrient treatments (n = 20 per trait per species; df = 1,18). Significant correlations at P < 0.05 (regular font) and P < 0.01 (bold font) are shown. (–) Denotes no data collected
 TreeN2-fixing treeShrubC3 grassC4 grass
Metrosideros polymorphaPsidium* cattleianumAcacia koaLeucaena* leucocephalaRubus hawaiiensisRubus* ellipticusDeschampsia nubigenaHolcus* lanatusEragrostis variabilisPaspalum* urvillei
Plant-level traits
 LWR0.570.73 0.50 0.46  −0.47 
 SWR     0.40
 RWR −0.51   −0.46  0.47 
 LAR−0.53 −0.65   −0.52 −0.79 
 R/S −0.53   −0.44  0.44 
Leaf-level traits
 Asat −0.80 0.810.530.620.690.54 0.59
 WUE0.65        0.67
 Narea0.660.83 0.800.520.650.620.65 0.51
 Nmass   0.48   0.57−0.52 
 LMA0.610.640.540.52  0.69 0.800.46
 Chl a/b  0.56    0.740.64
 Chl/N −0.64      −0.63 
 Vmax0.60  0.750.66  0.75
 ΦPSII   0.790.740.720.750.61 0.71
 Fv/Fm  −0.66     −0.56 


plasticity in low resource environments

High phenotypic plasticity has long been thought to be a characteristic of invasive species (Baker 1965; Marshall & Jain 1968) but few studies have properly evaluated this claim by comparing plasticity among invasives and phylogenetically related natives or introduced non-invasive species (Williams & Black 1994; Gerlach & Rice 2003; Burns & Winn 2006). In a recent review of the existing literature, Richards et al. (2006) concluded that the ability to capitalize on increased resource availability is a prevalent trait among invasive species. This conclusion concurs with predictions of the fluctuating resource hypothesis (Davis et al. 2000), which states that a community becomes more susceptible to invasion when resource availability is increased. Thus, as habitats experience alterations in the availability and timing of resources due to changing climate and anthropogenic disturbance, increased resource acquisition by invaders may alter competitive outcomes between invaders and natives. But should we extend this theory to low resource systems where species are thought to display low trait plasticity?

Working in a nutrient-poor environment, I found the magnitude of plasticity for leaf-level (PIV = 0.06–0.57) and plant-level (PIV = 0.19–0.66) traits across a suite of native and invasive species to be comparable to that found in species occupying higher resource habitats. In response to variation in light availability, Niinemets et al. (2003) found a PIV range of 0.18–0.59 for leaf-level traits in two evergreen shrubs occurring in a nutrient-rich site in Belgium (30–45 kg N ha−1 yr−1 via atmospheric deposition). Other studies in low resource systems, including a phosphorus-deficient, arid Mediterranean systems, found similar plasticity ranges for a diverse set of leaf- and plant-level traits (PIV = 0–0.80; Balaguer et al. 2001). The observation of considerable trait plasticity in species adapted to nutrient-poor environments runs counter to the argument that plasticity is too costly in stressful environments (e.g. Grime et al. 1986). However, the degree of plasticity in any environment will depend on the traits being examined. In a series of studies, Valladares et al. (2000, 2002) found that shade-tolerant species have low physiological plasticity and high morphological plasticity, while the reverse is true for shade-intolerant species.

species differences in trait plasticity

Invaders generally displayed higher trait plasticity than natives in response to altered nutrient availability while trait plasticity did not differ among native and invasive species in response to light availability. As all five pairs occur in nitrogen-poor soils (while only two of the five pairs also occur in light-limited forests), invaders in these habitats may benefit from the ability to respond to increased nutrient availability. Invasive species displayed trait plasticity consistent with success in nutrient-rich habitats. First, invasive species increased above-ground allocation (lower RWR, R/S) under high nutrient conditions, which concurs with expectations that plants will optimally partition biomass to maximize the capture of limiting resources (Bloom et al. 1985). Second, invasive species responded to increased nutrient availability with increased leaf N content (Narea) and photochemical function (Vmax, ΦPSII). Over time, these changes may lead to increased photosynthetic rate and growth under high nutrient conditions, but this effect was not observed here.

As a consequence of indeterminate growth, plants are inherently plastic; but not all trait plasticity is adaptive (Sultan 1995). Thus, for trait plasticity to contribute to plant invasiveness, it must be correlated with plant performance. As I used biomass instead of reproductive measures to assess plant performance, my data do not directly address adaptive plasticity. However, correlations of traits with biomass across the environmental gradient suggested that several plant- and leaf-level traits correlate strongly with plant fitness.

While light use efficiency (ΦPSII) correlated strongly with biomass in two native and three invasive species, the functional significance of several other traits varied among the two species groups. Asat and Narea showed strong correlations with biomass in four invasive but only one native species, while LMA correlated strongly with biomass in three native but only one invasive species. However, a complete understanding of the functional significance of trait plasticity may be especially difficult in low resource environments where plants display long leaf lifespan, high concentrations of defense compounds, low tissue nutrient content, and thicker leaves, all of which may result in reduced rates of photosynthesis and growth (Coley et al. 1985). While advantageous on long timescales (e.g. generation), these resource conservation traits (e.g. WUE, PNUE, LMA, leaf lifespan) may not correlate with fitness measures on short timescales (e.g. 1-year study), which may explain the insignificant correlations between many traits and plant biomass observed here.

While a few native species maintained thick leaves (LMA) and high light use efficiency (ΦPSII) relative to paired invasive species, native species as a group did not display higher resource-use efficiency than invasive species. These data support the conclusion from a recent field survey that invasive species may succeed in low resource systems by employing resource conservation traits similar to the native species adapted to these systems (Funk & Vitousek 2007). The observation that invaders displayed high trait plasticity and resource conservation contradicts the idea that there exists a trade-off between these two strategies (Grime et al. 1986). In fact, I observed no correlation among species PIV and any measure of resource-use efficiency. However, this assessment was limited to just ten species and does not provide a robust examination of this trade-off.

Several traits which showed low plasticity (e.g. Fv/Fm, Chl a/b) correlated strongly (P < 0.01) with biomass in some species while other traits (e.g. RWR, R/S, Nmass) showed high plasticity but did not correlate strongly with biomass. Although I did not directly assess adaptive plasticity, these data support the idea that small differences in the plasticity of some traits can have profound fitness consequences (Givnish 2002; Heschel et al. 2004; Funk et al. 2007). This result highlights the importance of exploring the functional significance of traits (rather than simply quantifying the amount of plasticity) in both genotypic- and species-level studies.

In conclusion, my data support the general paradigm that invasive species display high trait plasticity. The few studies that previously examined trait plasticity in invasive and native or non-invasive species in low resource environments found mixed support for higher trait plasticity in invasives (Williams & Black 1994; Padgett & Allen 1999; Weber & D’Antonio 1999; Cordell et al. 2002). However, this uncertainty may stem from the different types of traits (morphological, physiological, reproductive) measured across studies. As plant traits vary substantially in PIV (0.06–0.66, this study), studies examining only a handful of leaf-level or plant-level traits may not accurately represent differences in plasticity among native and invasive species.

Differences in phenotypic plasticity between native and invasive species will influence how these species will respond to changing environmental conditions. While low resource systems generally experience low levels of invasion (Burke & Grime 1996; Daehler 2003), these systems may become more susceptible to invasion as environmental conditions change and propagule pressure increases. Invasive species displayed high plasticity for traits (Narea, Vmax, ΦPSII) that correlated strongly with biomass across several species, which suggests that invaders in these low resource systems will benefit from N enrichment. Future work should focus on understanding plasticity in environments where plant productivity is limited by multiple interacting factors. As a large proportion of low resource environments are water limited, the influence of water availability on plasticity may be particularly important.


Author thanks R. Montgomery, S. McDaniel, N. Kuamo’o, H. Farrington, D. Turner, S. Dale, S. Hall, R. Martin, R. Schneider, S. Madden and D. Benitez for field and laboratory support and P. Vitousek, M. Rejmánek, C. Richards, N. Muth, L. Sack, P. Alpert and the Vitousek laboratory for discussion and comments on the manuscript. This research was made possible as a result of the National Parks Ecological Research Fellowship Program, a partnership between the National Park Service, the Ecological Society of America and the National Park Foundation. It is funded by a generous grant from the Andrew W. Mellon Foundation.