The evolution of increased competitive ability, innate competitive advantages, and novel biochemical weapons act in concert for a tropical invader

Authors

  • Rui-Min Qin,

    1. Key Laboratory of Tropical Forest Ecology, Xishuangbanna Tropical Botanical Garden, Chinese Academy of Sciences, Kunming, China
    2. Graduate University, Chinese Academy of Sciences, Beijing, China
    Search for more papers by this author
    • These authors contributed equally to this work.
  • Yu-Long Zheng,

    1. Key Laboratory of Tropical Forest Ecology, Xishuangbanna Tropical Botanical Garden, Chinese Academy of Sciences, Kunming, China
    2. Graduate University, Chinese Academy of Sciences, Beijing, China
    Search for more papers by this author
    • These authors contributed equally to this work.
  • Alfonso Valiente-Banuet,

    1. Instituto de Ecología, Departamento de Ecologôa de la Biodiversidad, Universidad Nacional Autônoma de Mexico, México D.F, México
    Search for more papers by this author
  • Ragan M. Callaway,

    1. Division of Biological Sciences, University of Montana, Missoula, MT, USA
    Search for more papers by this author
  • Gregor F. Barclay,

    1. Department of Life Sciences, University of the West Indies, St. Augustine, Trinidad and Tobago
    Search for more papers by this author
  • Carlos Silva Pereyra,

    1. Instituto de Ecología, Departamento de Ecologôa de la Biodiversidad, Universidad Nacional Autônoma de Mexico, México D.F, México
    Search for more papers by this author
  • Yu-Long Feng

    Corresponding author
    1. College of Bioscience and Biotechnology, Shenyang Agricultural University, Shenyang, Liaoning Province, China
    • Key Laboratory of Tropical Forest Ecology, Xishuangbanna Tropical Botanical Garden, Chinese Academy of Sciences, Kunming, China
    Search for more papers by this author

Author for correspondence:

Yu-Long Feng

Tel: +86 24 88487163

Email: yl_feng@tom.com; fyl@xtbg.ac.cn

Summary

  • There are many non-mutually exclusive mechanisms for exotic invasions but few studies have concurrently tested more than one hypothesis for the same species.
  • Here, we tested the evolution of increased competitive ability (EICA) hypothesis in two common garden experiments in which Chromolaena odorata plants originating from native and nonnative ranges were grown in competition with natives from each range, and the novel weapons hypothesis in laboratory experiments with leachates from C. odorata.
  • Compared with conspecifics originating from the native range, C. odorata plants from the nonnative range were stronger competitors at high nutrient concentrations in the nonnative range in China and experienced far more herbivore damage in the native range in Mexico. In both China and Mexico, C. odorata was more suppressed by species native to Mexico than by species native to China. Species native to China were much more inhibited by leaf extracts from C. odorata than species from Mexico, and this difference in allelopathic effects may provide a possible explanation for the biogeographic differences in competitive ability.
  • Our results indicate that EICA, innate competitive advantages, and novel biochemical weapons may act in concert to promote invasion by C. odorata, and emphasize the importance of exploring multiple, non-mutually exclusive mechanisms for invasions.

Introduction

Successful biological invasions depend to a large degree on interactions between introduced organisms and species native to the invaded systems. For example, invasive plants often outcompete native plants (Garcia-Serrana et al., 2007; Mangla et al., 2011; Callaway et al., 2012), and this stronger competitive ability has been associated with higher resource capture and utilization efficiency (Feng et al., 2007), stronger allelopathic effects (Callaway & Aschehoug, 2000; Ni et al., 2010), and release from consumer pressure (Keane & Crawley, 2002). Furthermore, the intensity of these interactions may be affected by their novelty in the invaded system (Hierro et al., 2005; Lamarque et al., 2011), or evolve in response to conditions in the nonnative range (Siemann & Rogers, 2001; Ridenour et al., 2008; Lei et al., 2011). The ecological effects of inherent novelty versus evolution on invasions are not likely to be mutually exclusive, but few studies have explored these ideas together or have studied multiple mechanisms affecting exotic invasion in general (e.g. DeWalt et al., 2004; Hierro et al., 2005; Williams et al., 2010).

Evolutionary responses to biota in the nonnative ranges of invaders have often been explored in the context of the ‘evolution of increased competitive ability’ (EICA) hypothesis. The EICA hypothesis posits that the release of exotic species from consumer pressure in their nonnative ranges can lead to decreased allocation to defense and a concomitant increase in allocation to growth and by extension to increased competitive ability (Blossey and Nötzold, 1995). A number of experiments have supported one or both predictions of the EICA hypothesis for some species (Siemann & Rogers, 2001; Bossdorf et al., 2004; Maron et al., 2004; Wolfe et al., 2004; Ridenour et al., 2008), but for other species no evidence was found for EICA (Vilà et al., 2003; Bossdorf et al., 2004). Feng et al. (2009, 2011) reported a mechanism for the EICA hypothesis. They found evolutionary shifts in nitrogen allocation from cell walls (defense) to photosynthesis (growth) in nonnative populations of an invasive species, resulting in faster growth and poorer structural defenses.

Competitive advantages of invaders over natives have been studied recently in the context of the novel weapons hypothesis (NWH), the idea that constitutive and induced biochemicals that are new to the nonnative ranges of the invader provide disproportionate allelopathic advantages against naïve native plant species (Callaway & Aschehoug, 2000; Inderjit et al., 2011a), against native soil biota through antibiotic effects (Callaway et al., 2008), or against herbivores (Lankau et al., 2004; Schaffner et al., 2011). For example, Callaway et al. (2012) found that the Eurasian Acroptilon repens occurred in denser stands with far lower relative abundances of natives in North America than in its native Uzbekistan, and Ni et al. (2010) found that this invader also exerted stronger competitive and allelopathic effects on North American species than on species from Uzbekistan. Similar results have been reported for the allelopathic effects of other invasives, including Ageratina adenophora (Inderjit et al., 2011b), Centaurea stoebe (Thorpe et al., 2009), Centaurea diffusa (Callaway & Aschehoug, 2000), Foeniculum vulgare (Colvin & Gliessman, 2011) and Prosopis juliflora (Kaur et al. 2012), for a suite of invasive species in Korea (Kim & Lee, 2011), and in a meta-analysis of invasive tree species (Lamarque et al., 2011).

Chromolaena odorata (L.) R. M. King and H. Robinson (Compositae) is native to North, Central, and South America but is a noxious invasive perennial herb or subshrub throughout much of Asia, Oceania, and Africa. It was first introduced into India as an ornamental plant in the middle of the 19th Century, and now has become one of the most invasive species in southern China. Chromolaena odorata often forms dense monocultures in habitats such as croplands, plantations, pastures, disturbed forests, roadsides, and riverbanks, causing great economic loss and threatening biodiversity (Zhang & Feng, 2007). More than 200 arthropod enemies have been found to attack C. odorata in its native range, a quarter of which are specialists (Zhang & Feng, 2007), but very few generalist or specialist enemies occur in the nonnative ranges of the invader. In this study, we tested the EICA hypothesis and the NWH by comparing the performances of C. odorata plants collected as seeds from different populations in the native and nonnative ranges and then grown in common gardens in Mexico and China. In this context, we compared interactions between C. odorata and Eupatorium japonicum, Eupatorium stoechadosmum, and Eupatorium heterophyllum, which are native to China, and Eupatorium ligudtrinum, which is native to Mexico. We chose Eupatorium species because the genus is very closely related to the genus ChromolaenaC. odorata is synonymous with Eupatorium odoratum. Resource supply can affect the outcomes of competition among natives and invaders (Besaw et al., 2011) and increased soil fertility has been shown to facilitate invasion of alien species in general (Daehler, 2003) and C. odorata specifically (Wang & Feng, 2005). Therefore we conducted the experiment in China at ambient and elevated soil nutrient concentrations. Finally, to explore the NWH we conducted an experiment with leachates made from C. odorata leaves and applied to seeds of eight species native to Mexico and seven species native to China. We asked whether C. odorata plants from populations in the nonnative range differ in size, competitive ability, and herbivore defense from plants from the native range; and whether species from the nonnative range of C. odorata differ in susceptibility to the effects of C. odorata leaf leachates from species from the native range of C. odorata.

Materials and Methods

Study sites and materials

A common garden pot experiment and germination experiment were conducted within the nonnative range of Chromolaena odorata (L.) R. M. King and H. Robinson at the Xishuangbanna Tropical Botanical Garden (21°560′ N, 101°150′ E; 570 m altitude) of the Chinese Academy of Sciences located in Mengla County, Yunnan Province, southwest China. An outdoor common garden experiment was conducted within the native range in Tlayacapan, Morelos, Mexico (18°57′ N, 98°58′ W; 1634 m altitude). Chromolaena odorata grew wild around the Mexican garden. In Xishuangbanna, the annual average temperature is 21.7°C; the mean temperature of the hottest month (July) is 25.3°C and that of the coolest month (January) is 15.6°C. The annual average precipitation is 1557 mm with a dry period from November to April. In Tlayacapan, the mean annual temperature is 19.3°C; the mean temperature of the hottest month (June) is 22.9°C and that of the coolest month (January) is 16.9°C. The mean annual precipitation is 988.8 mm with a dry period from November to April (García, 1988).

For both common garden experiments, we collected seeds of C. odorata from six populations in its native range (five from Mexico and one from Trinidad; Supporting Information Table S1) and six populations in its nonnative range (three from China, two from Vietnam, and one from Laos) in 2009. We also collected seeds from native Eupatorium lingustrinum DC in Mexico and from native Eupatorium japonicum Thunb., Eupatorium stoechadosmum Hance, and Eupatorium heterophyllum DC. in China. All of these species were sympatric, ecologically similar, and phylogenetically related to C. odorata (syn. E. odoratum). For each species, seeds were collected from more than 10 individuals that were at least 20 m apart from one another. Seeds from individual plants were not mixed and thus seeds collected from each individual comprised a ‘seed family’.

Common garden pot experiment in China

Seeds of the Chinese natives E. japonicum and E. stoechadosmum, the Mexican native E. lingustrinum, three C. odorata populations from the invasive range, and four C. odorata populations from the native range (Table S1) were sown in seedling raising trays (128 36-ml cells per tray) in a glasshouse in November 2009. We originally chose five populations randomly from each regional collection, but seeds from three of the ten populations did not germinate. In February 2010, when the seedlings were c. 5 cm tall, similar-sized vigorous seedlings were transplanted into 15-dm3 pots. Ten individuals from each C. odorata population (two per seed family and five seed families per population) and 10 individuals for each Eupatorium species were planted alone for a total n of 100. We also transplanted 10 individuals of each of the seven C. odorata populations into pots with each of the three Eupatorium species, with plants 10 cm apart, for a total n of 210. Pots contained a mixture of 70% forest topsoil and 30% river sand. Topsoil was used to provide a natural supply of macro- and micronutrients and the river sand provided a texture with adequate drainage and facilitated the harvesting of fine roots. All seedlings were grown in shade with 50% irradiance for 4 wk to facilitate initial survival and then were grown in full sun. Seedlings were divided into two groups; one was treated with compound fertilizer (nitrogen: phosphorus: potassium 15:15:15) monthly from April to July at the rate of 1 g per pot (10 kg soil), and the other group was the low-nutrient control. Five pots per population of C. odorata (one from each seed family) or native species for each nutrient and each competition (alone or with neighbor) treatment were established. Pots were assigned to five blocks in the common garden, and each block included 62 pots, one pot for each population or species × nutrient × competition treatment. The seedlings were watered daily with a drip irrigation system. Pots were weeded when necessary and no pesticides were used.

Plants did not flower and in August 2010 all plants (including roots) were harvested, oven-dried at 60°C for 48 h, and weighed. To evaluate competitive ability, the competitive ‘response’ of each species was measured as the per cent change in performance (biomass) of the species when grown with competition, that is, ((PcompPsingle)/Psingle) × 100, where Psingle is plant performance when grown without competition and Pcomp is plant performance when grown with interspecific competition. The competitive ‘effect’ of each species was also measured as the per cent change in the performance of its competitor. In this study, Psingle was the average of all replicates per species or population per treatment and Pcomp was the value of the individual replicate.

Common garden experiment in Mexico

Seeds of E. heterophyllum and E. stoechadosmum from China and E. lingustrinum from Mexico and seeds from each of the 10 seed families of each of the five C. odorata populations from each range (randomly selected; Table S1) were sown into a seed bed located in a glasshouse in October 2009. In January 2010, when the seedlings were c. 10 cm tall, similar-sized vigorous seedlings were transplanted into an outdoor garden. We grew 10 individuals of each C. odorata population (one from each seed family) and Eupatorium species alone, and 10 individuals of each C. odorata population with each of the three Eupatorium species (6 cm apart from each other). All individuals and competing pairs were 60 cm from any other plant or pair and were assigned to 10 blocks in the common garden, 13 individuals (one from each of the 10 populations of C. odorata and the three Eupatorium species) and 30 competing pairs (each C. odorata population with each of the three Eupatorium species) per block. Individuals and competing pairs were arranged randomly in each block. Seedlings were watered every other day at the rate of 2000 ml per seedling or seedling pair in the dry season (January–May). Plots were weeded and no pesticides were used.

In October 2010, five individuals (without competitors) per population and one branch per individual (with ~50 leaves) were randomly selected for measurement of leaf herbivory. The number of damaged leaves and the total number of leaves were counted for each individual, and the percentage of damaged leaves was calculated as (the number of damaged leaves/the total number of leaves) × 100%. Total leaf area and the area damaged by enemies were visually estimated by comparison with a paper square of 10 cm by 10 cm dimensions (accurate to 0.1 cm) for each leaf, and the percentage of leaf area loss was calculated for each damaged leaf and each sample individual.

Plants did not flower, and in November 2010 the aboveground parts of all plants were harvested, oven-dried at 60°C for 48 h, and weighed. Competitive response and effect were calculated as described in the section of the Common garden pot experiment in China.

Leachate experiment

We compared the effects of C. odorata leaf leachate on the germination and growth of species native to Mexico with the effects on species native to China. Seeds were collected in 2010 from seven native species from China (Carex baccans Nees, Eupatorium japonicum, Eupatorium fortunei Turcz, E. stoechadosmum, Polygonum molle D. Don, Vernonia cinerea (Linn.) Less, and Vernonia volkameriifolia (Wall.) DC), and eight native species from Mexico (Aldama dentata La Llave, Calea ternifolia Kunth, Cosmos sulphureus Cav, Lagascea rigida (H. B. K.) Stuessy, Lasianthaea helianthoides DC. var. helianthoides, Tagetes tenuifolia Cav., E. ligustrinum, and Eupatorium sp.). All native species from each range were sympatric with C. odorata. In January 2011, fully expanded leaves were collected from more than 10 C. odorata plants growing wild in the grounds of Xishuangbanna Tropical Botanical Garden, dried in the shade at room temperature and powdered. Leaf powder was soaked in distilled water (2.5 g of leaf powder per 100 ml of distilled water) at 4°C for 48 h, and filtered twice through double layers of gauze. The filtrate, which represented a 2.5% leaf extract, was kept at 4°C until used.

Seeds were soaked in 0.3% aqueous potassium permanganate solution for 15 min to sterilize them, washed twice with distilled water, and placed on two layers of filter paper in 12-cm Petri dishes, 30 seeds per dish. Four ml of leaf extract of C. odorata was added to each dish. Four concentrations of leaf extract (2.5%, 1.25%, 0.25% and 0.0% (distilled water as a control)) were used, with five replicates (Petri dishes) per species per treatment. The number of germinated seeds (emergence of the radicle) was recorded daily. Radicle length was measured when no further germination of seeds occurred for three consecutive days. The germination rate was calculated as (the number of germinated seeds/30 (total number of seeds per Petri dish)) × 100%. The germination rate or radicle length measured for the control was significantly different between species. The relative values of these variables (per cent of control) were calculated for each species and treatment to facilitate interspecific comparisons.

Statistical analyses

The effects of range, population nested within range, nutrient, interaction of range and nutrient, and interaction of population nested within range and nutrient on variables measured in the common garden experiment in China were analyzed using three-way nested ANOVA, with all the factors as fixed factors (univariate of general linear model; Table S2). The significance of differences between plants from invasive and native populations of C. odorata at each nutrient concentration in China and in Mexico was determined using two-way nested ANOVAs, with range and population nested within range as fixed factors (Figs 1, 3, 5). In this study, population nested within range was treated as a fixed factor because of the low number of populations (Siemer & Joormann, 2003). In Mexico, but not in China, Eupatorium species native to Mexico suppressed C. odorata plants from both ranges more than Eupatorium species native to China did (Fig. S2). Thus, we also tested intraspecific differences (between C. odorata plants from invasive and native populations) in the competitive responses to the natives from China and Mexico separately. Neither the altitude nor the latitude at which seeds of C. odorata were collected was used as a covariate, as neither of them was significantly correlated with the performance of C. odorata plants (> 0.05). The significance of differences in variables measured in both common garden experiments among invasive and native populations of C. odorata, species native to China, and species native to Mexico (Figs 2, 4), and the significance of differences in relative germination rate and relative radicle length between species native to China and Mexico at each leaf extract concentration of C. odorata (Figs 6, S3) were determined using ANOVAs. All analyses were conducted using spss 13.0 (SPSS Inc. Chicago, IL, USA).

Results

Common garden pot experiment in China

When grown without competition, C. odorata plants from the native range were larger than plants from the nonnative range at high nutrient supply (Fig. 1a). Total biomass decreased significantly with decreasing nutrient supply (Fig. 1a,b; Table S2), and at low nutrient supply, total biomass did not differ between plants from the two ranges (Fig. 1b). Competition from all Eupatorium species decreased the total biomass of C. odorata plants from both ranges at both nutrient concentrations. But, importantly, at high nutrient supply C. odorata plants originating from the nonnative range in China had superior competitive responses to the Eupatorium species compared with C. odorata plants from the native range in Mexico (Fig. 1c). In other words, competition-driven decreases in total biomass were significantly greater for C. odorata plants from the native range than for C. odorata plants from the nonnative range. This did not occur at low nutrient supply (Fig. 1d), which was consistent with the significant interaction between range and nutrient supply (Table S2).

Figure 1.

Total biomass for Chromolaena odorata plants from invasive (open bars) and native (closed bars) populations grown without competition at high (a) and low (b) nutrient concentrations in the common garden in China, and competitive response to three Eupatorium species (two natives to China and one native to Mexico; per cent changes in total biomass) at high (c) and low (d) nutrient concentrations. Narrow bars indicate mean +SE (= 5 for monoculture; = 15 for competition) for each population; two thicker bars in the center depict mean +SE for each range using the mean of each population as replicates. Significant differences between ranges according to nested ANOVAs (Supporting Information Table S3) at: *, < 0.05; ***, < 0.001.

At high nutrient supply, C. odorata plants from both the nonnative and native ranges were much less suppressed by the native species from China than the native species from China were suppressed by C. odorata (Figs 2c, S1a). In other words, C. odorata plants from both ranges had superior competitive responses and stronger competitive effects than the native species from China when these species were interacting. In the low-nutrient treatment, the competitive advantage of the invader over the Eupatorium species native to China decreased or even disappeared (Figs 2d, S1b). The two native Eupatorium species from the invaded range in China were much more suppressed by C. odorata than were the Eupatorium species from the native range in Mexico in the high-nutrient treatment (Fig. 2c); that is, the native species from Mexico had a superior competitive response to C. odorata compared with natives from China. In the low-nutrient treatment, however, the competitive effects of C. odorata on E. stoechadosmum native to China and E. ligustrinum native to Mexico were similar (Fig. 2d).

Figure 2.

Total biomass for plants grown without competition (a, b), and change in this variable caused by competition (competitive response or effect) (c, d) for invasive (open bars) and native (closed bars) populations of Chromolaena odorata, Eupatorium stoechadosmum (vertically striped bars; native to China), Eupatorium japonicum (hatched bars; native to China), and Eupatorium ligustrinum (horizontally striped bars; native to Mexico) grown at high (a, c) and low (b, d) nutrient concentrations in the common garden in China. For C. odorata, change in total biomass was caused only by natives to China. Means +SE are shown (for monoculture, = 5; for competition, = 30 for invasive populations of C. odorata,= 40 for native populations of C. odorata, and = 35 for natives). Different letters indicate significant differences among species at < 0.05 according to ANOVAs (LSD test, Table S4).

Common garden experiment in Mexico

Without competitors, C. odorata plants derived from populations in the native range did not differ in aboveground mass from plants from the nonnative ranges (Fig. 3a). Furthermore, the competitive responses of C. odorata to Eupatorium species did not differ between C. odorata plants from invasive and native populations, whether tested using the responses to all natives together (Fig. 3b) or to natives from different countries separately (Fig. S2).

Figure 3.

Aboveground mass (a) for Chromolaena odorata plants from invasive (open bars) and native (closed bars) populations grown without competition in the common garden in Mexico, and competitive response to three Eupatorium species, two native to China and one native to Mexico (per cent change in aboveground mass) (b). Narrow bars indicate mean +SE (= 10 without competition; = 20 with competition) for each population; two thicker bars in the center depict mean +SE for each range using the mean of each population as replicates. The differences between ranges were not significant according to nested ANOVAs (Table S5).

Generally consistent with the results from the pot experiment in China (Fig. 2), C. odorata plants from both ranges were much less suppressed by the Eupatorium species native to China than these were suppressed by C. odorata (Figs 4, S1c). In other words, C. odorata plants from both ranges had superior competitive responses and stronger competitive effects compared with the native species from China. Most importantly, the native species from Mexico had a superior competitive response to C. odorata and a stronger competitive effect on C. odorata compared with the natives from China. The two native Eupatorium species from China were much more suppressed by C. odorata than the Eupatorium species from Mexico (Fig. 4), and C. odorata was much more suppressed by the native Eupatorium species from Mexico than by the two Eupatorium species native to China (Fig. S2).

Figure 4.

Aboveground mass (a) for plants grown without competition in the common garden in Mexico, and change in this variable caused by competition (competitive response or effect) (b) for invasive (open bars) and native (closed bars) populations of Chromolaena odorata, Eupatorium stoechadosmum (vertically striped bars; native to China), Eupatorium heterophyllum (diagonally striped bars; native to China), and Eupatorium ligustrinum (horizontally striped bars; native to Mexico). For C. odorata, change in aboveground mass were caused by natives only from China. Means +SE are shown (for monoculture, = 10; for competition, = 100 for native species from Mexico, and = 50 for others). Different letters indicate significant differences among plants at  0.05 according to ANOVAs (LSD test; Table S6).

Herbivores attacked C. odorata plants from populations from the nonnative ranges in Southeast Asia more, and did more damage to them, than plants derived from the native range. The number of damaged leaves was 107.9% higher for plants from Southeast Asian populations of C. odorata than for plants from native populations, and plants from the nonnative ranges experienced 174.9% more leaf area loss than plants from the native range (Fig. 5).

Figure 5.

Differences in damaged leaves per plant (a) and damaged leaf area per plant (b) between Chromolaena odorata plants from invasive (open bars) and native (closed bars) populations grown without competition in the common garden in Mexico. Narrow bars indicate mean +SE (= 5) for each population; two thicker bars in the center depict mean +SE for each range using the mean of each population as replicates. Significant differences between ranges according to nested ANOVAs (Table S6): ***, < 0.001.

Leachate experiment

Leaf extracts from C. odorata suppressed the mean relative germination rates (as percentages of the rates for the controls) of species native to China far more than those of species native to Mexico (also the native range of C. odorata) at all concentrations tested (Figs 6a, S3). For comparison, inhibition proportions were 65.4% and 44.9% higher for species native to China than those for species native to Mexico at 1.25% and 2.5% concentrations, respectively. At the 0.25% leachate concentration, leaf extracts from C. odorata inhibited germination by 59% for species native to China but had no effect on species native to Mexico. The 0.25, 1.25, and 2.5% leaf extracts significantly inhibited seed germination for one (12.5% of the eight species), four (50.0%), and seven (87.5%) of the eight species from Mexico, respectively; but for four (57.1% of the seven species), seven (100%), and seven (100%) of the seven species from China (Fig. S3). The lowest leaf extract concentration (0.25%) significantly promoted seed germination (relative germination rate > 100%) for two species from Mexico but no species from China.

Figure 6.

Relative (per cent of control) germination rate (a) and relative radicle length (b) for native species from China (= 7; open circles) and Mexico (= 8; closed circles). Means ±SE are shown for each range using the mean of each species as replicate; = 5 for each species. The differences in the two variables between species native to China and Mexico were significant for all treatments ( 0.001) according to ANOVAs (Fig. S3).

Similarly, the leaf extracts from C. odorata also inhibited seedling growth of the species native to China far more than that of species native to Mexico (Figs 6b, S3). The mean values of relative radicle length of the species native to Mexico were much higher than those of the species native to China at 0.25% (62.3% vs 12.7% of the control, respectively) and 1.25% (25.2% vs 1.4% of the control) concentrations. At the highest extract concentration (2.5%), the relative radicle length of three species native to Mexico ranged from 22 to 43% of the control, whereas no species native to China showed measurable growth (Fig. S3).

Discussion

Our results link apparent rapid evolutionary changes in competitive ability and herbivore defense occurring in the nonnative range to inherently disproportionate competitive and allelopathic advantages of the invader over native species. The latter may derive from long-term evolutionary relationships among members of a community. Thus, we found evidence for both EICA and the NWH. Chromolaena odorata had stronger competitive effects on Eupatorium species from China (the invaded range) than on Eupatorium species from Mexico (the native range of C. odorata) in two different experiments. These competitive effects did not correspond with the biomass produced by the competing species in China, and the germination and growth of species native to China were far more inhibited by extracts from C. odorata leaves than those of species native to Mexico. This is consistent with a growing body of evidence for benefits that some invasive species gain from the production of biochemicals to which species native to the nonnative ranges of invaders are not adapted, perhaps because they lack a common evolutionary history (Callaway & Aschehoug, 2000; Ni et al., 2010; Colvin & Gliessman, 2011; Inderjit et al., 2011b; Kim & Lee, 2011; Lamarque et al., 2011; Kaur et al., 2012). These biogeographic differences in allelopathic effects also provide one possible explanation, and potential integration of the effects of EICA and the NWH, for the stronger competitive effects of C. odorata on species from China than on species from Mexico. Allelochemicals released from C. odorata plants and residues have been shown to accumulate in soils and inhibit plant growth in the field in nonnative ranges (Onwugbuta-Enyi, 2001; Singh & Angiras, 2008); however, to our knowledge specific chemicals have not been identified and nor have the biogeographic patterns evident in our Petri dishes been explored under field conditions where allelopathic effects may be attenuated.

At the high nutrient concentration in China, C. odorata plants from invasive ranges demonstrated superior competitive responses compared with C. odorata plants from native ranges. This corresponded with other evidence that C. odorata plants from China were more poorly defended than plants from Mexico, and both are consistent with the hypothesis that the stronger competitive ability of some invasive species may derive from evolving to decrease allocation to costly structural and chemical defenses (Blossey & Nötzold, 1995; Feng et al., 2009). However, these competitive advantages were not associated with greater biomass or growth rates as originally proposed. When grown without competition, C. odorata plants from nonnative-range populations were actually 14.2% smaller than plants from native-range populations. It was only in the presence of competitors that plants from the nonnative range demonstrated a superior response compared with conspecifics from the native range, with Eupatorium competitors eliciting a 15.1% decline in total mass for the nonnative-range C. odorata vs 34.9% for native-range C. odorata. The increased interspecific competitive ability of C. odorata plants from nonnative ranges may be associated with directional selection for genotypes that produce higher amounts of allelochemicals, such as indicated for the invasive Centaurea maculosa (Ridenour et al., 2008).

Our results are consistent with other research reporting superior competitive ability for invasive populations of exotic species relative to co-occurring natives (Vilà & Weiner, 2004; Werner et al., 2010). Our results are also consistent with evidence that the evolution of increased competitive ability after introduction contributes to the competitive advantage of C. odorata and other invasive species (Ridenour et al., 2008; Barney et al., 2009). However, to our knowledge no previous studies have explored the importance of the evolution of increased competitive ability and innate competitive advantages in a parallel study. Our results indicate that innate competitive advantages contribute greatly to the higher competitive ability for invasive populations of C. odorata compared with natives from China (Figs 2, 4, S1).

Because we used common garden experiments, differences among populations and ranges were not confounded by plasticity (Reznick & Ghalambor, 2001) but we did not exclude founder effects or maternal effects, as for most studies of the evolution of invasives. Performing experiments in more than one environment in each range increases the potential to detect evolutionary differences between native and nonnative populations of invasive species but this does not eliminate possible misinterpretation as a result of founder effects. Few other biogeographic studies have been conducted in more than one garden per range, and very few have compared responses in invasive and native ranges (but see Maron et al., 2007; Widmer et al., 2007; Williams et al., 2008). By establishing comparative biogeographic experiments in both ranges, our design incorporated the likelihood that population- or regional source-based differences in competitive ability may vary when measured in different abiotic conditions, as demonstrated by the present study (Table S2) and Maron et al. (2004). The overall differences in competitive ability and defense between C. odorata plants from native and nonnative populations indicate biogeographical differentiation, but to establish whether this differentiation is attributable to evolution after introduction, we must compare invasive populations with their specific source populations (e.g. Dlugosch & Parker, 2008), and source populations are rarely known with certainty. Based on internal transcribed spacer (ITS) sequences, Scott et al. (1998) and von Senger (2002) found significant genetic differentiation between invasive and native populations of C. odorata, and concluded that C. odorata populations in Asia, West Africa, and Australia originally originated from Trinidad. Evidence that C. odorata which invaded Asia may also originate from Trinidad was also obtained by X-Q. Yu et al. (unpublished) using nuclear (ITS) and chloroplast (intergenetic spacers between QB protein gene and photosystem II protein D1 gene (psbA-trnH) and between beta subunit gene of ATP synthase and large subunit gene of ribulose-1,5-bisphosphate carboxylase/oxygenase (atpB-rbcL)) DNA sequences. Here we found that C. odorata plants from the likely source population (Trinidad; the third of the four native populations from left to right in Fig. 1c) showed an inferior competitive response compared with C. odorata plants from the invasive populations, which further supports the likelihood that C. odorata in Asia has undergone evolutionary changes consistent with the EICA hypothesis.

In contrast to the results in the high nutrient supply treatment in the pot experiment in China, we did not find greater competitive ability of C. odorata from nonnative populations at low nutrient supply in China or in the common garden in Mexico. In part, these results are consistent with the advantages invaders often appear to gain with increasing resource availability (Daehler, 2003; Mangla et al., 2011; but see Garcia-Serrana et al., 2007). However, the absence of regional differences in Mexico may have occurred for several reasons, including unmeasured differences in abiotic and biotic environments between ranges (Maron et al., 2004). Furthermore, any possible novel biochemicals produced by the invader may not have been as important in the Mexican experiment which was conducted in the field, and thus in the context of soil biota and herbivores.

To evaluate competitive effects and responses, we used three Eupatorium species native to China and one Eupatorium native to Mexico in the experiments. These choices were intended to provide a reasonable set of phylogenetically controlled competitors but not to explicitly compare the effects and responses of Chinese and Mexican sources of Eupatorium to C. odorata from different regions. However, the results from these comparisons were consistent with the results from the experiment with leaf extracts, and we found that the Eupatorium species from Mexico was a far better competitor than the native Chinese Eupatorium species. For example, competition from C. odorata plants caused 39.9% and 64.5% lower decreases in aboveground mass for the Eupatorium species from Mexico than for the Eupatorium species from China when evaluated in China (R. Qin, unpublished data) and Mexico (Fig. 4), respectively. In addition, the Mexican Eupatorium also had a much greater competitive effect on C. odorata plants from both ranges than the Chinese Eupatorium species when evaluated in Mexico (Fig. S2). To our knowledge, only one other study has examined the EICA hypothesis using phylogenetically controlled competitors from the native and nonnative ranges. McKenney et al. (2007) compared the growth of different genotypes of Lepidium draba from its native European and introduced western USA ranges in competition with the North American Festuca idahoensis and the European Festuca ovina. Contrary to our results, they found no differences in the performances of L. draba from the different ranges under any conditions, but corresponding to our results, F. ovina suppressed the growth of L. draba much more than that of F. idahoensis.

In conclusion, C. odorata plants from the native range demonstrated superior competitive ability against species native to the range it invaded, showing innate competitive advantages. Also, native species from the invasive range in China were more vulnerable to allelochemicals presumably present in the leachate made from C. odorata than natives from the native range in Mexico, which is consistent with the NWH. Plants from invasive populations of C. odorata demonstrated a superior competitive ability and inferior defensive ability against natural enemies compared with plants from native populations, which is consistent with the EICA hypothesis. In Mexico, however, and in the low-nutrient treatment in China, plants from the invasive range in Asia did not demonstrate superior competitive ability indicating substantial biotic and abiotic conditionality for our results and suggesting that different environments should be considered when testing mechanisms underlying biological invasions. Our results emphasize empirically that the different mechanisms that drive invasions are not mutually exclusive.

Acknowledgements

This study was funded by projects of the National Natural Science Foundation of China (30830027; 31270582) and Knowledge Innovation Program of Chinese Academy of Sciences (KSCX2-YW-Z-1019).

Ancillary