Differential responses to fertilization and competition among invasive, noninvasive alien, and native Bidens species

Abstract Comparative studies of invasive, noninvasive alien, and native congenic plant species can identify plant traits that drive invasiveness. In particular, functional traits associated with rapid growth rate and high fecundity likely facilitate invasive success. As such traits often exhibit high phenotypic plasticity, characterizing plastic responses to anthropogenic environmental changes such as eutrophication and disturbance is important for predicting the invasive success of alien plant species in the future. Here, we compared trait expression and phenotypic plasticity at the species level among invasive, noninvasive alien, and native Bidens species. Plants were grown under nutrient addition and competition treatments, and their functional, morphological, and seed traits were examined. Invasive B. frondosa exhibited higher phenotypic plasticity in most measured traits than did the alien noninvasive B. pilosa or native B. bipinnata. However, differential plastic responses to environmental treatments rarely altered the rank of trait values among the three Bidens species, except for the number of inflorescences. The achene size of B. frondosa was larger, but its pappus length was shorter than that of B. pilosa. Two species demonstrated opposite plastic responses of pappus length to fertilization. These results suggest that the plasticity of functional traits does not significantly contribute to the invasive success of B. frondosa. The dispersal efficiency of B. frondosa is expected to be lower than that of B. pilosa, suggesting that long‐distance dispersal is likely not a critical factor in determining invasive success.

Although the contribution of such traits to invasive success depends on the ecological context (Catford et al., 2009), an investigation of trait variations between invasive and noninvasive plants is a first step in identifying candidate invasive alien species for weed-risk assessments and in formulating an trait-based predictions to invasiveness (Gallagher et al., 2014;van Kleunen et al., 2015;Mathakutha et al., 2019).
The comparison between native and invasive plant species is a widely adopted method to infer traits that determine invasive success. However, the information gleaned often seems inadequate to evaluate attributes of invasive species (van Kleunen et al., 2015). First, the choice of a target native species is critical, because a comparison between invasive and co-occurring native species with different life cycles is likely to yield misleading results (Daehler, 2003;Vilá & Weiner, 2004). In addition, a comparison between invasive and noninvasive alien species would be more appropriate to identify traits contributing to invasive success (van Kleunen, Dawson, et al., 2010). A comparative study considering invasive, noninvasive alien, and native plant species, all of which are congenic, is required to gain insight into the traits that determine invasiveness.
Traits associated with resource capture ability are highly plastic to environmental conditions, particularly resource availability (Davidson et al., 2011;Funk, 2008;Gioria & Osborne, 2014;Godoy et al., 2012). Invasive plants tend to exhibit higher phenotypic plasticity than noninvasive plants, contributing to their establishment in novel environments and competition with existing vegetation (Davidson et al., 2011). Given the ecological significance of plasticity in invasive success, phenotypic plasticity to anthropogenic environmental change is of particular interest for the prediction of the future dynamics of plant invasion. Human activities continue to increase disturbance and eutrophication in terrestrial habitats, both of which are thought to facilitate the invasion of alien plant species (Chytrý et al., 2008;Davis et al., 2000;Jauni et al., 2015).
Disturbance removes competing species, consequently decreasing biotic interactions between species while increasing soil nutrient availability. Though plant responses to nutrient availability are widely studied, little information is available on phenotypic plasticity to environmental disturbance (Gioria & Osborne, 2014).
Since plants are sessile organisms, seed dispersal and germination are critical life stages for habitat selection and thereby influence species range (Donohue et al., 2010). A widely held hypothesis suggests that invasive plants tend to have smaller fruits or seeds than native species, thus facilitating seed dispersal and rapid range expansion (Coutts et al., 2011;Rejmánek & Richardson, 1996).
However, diverse fruit traits other than seed size can affect dispersal efficiency (Römermann et al., 2005), and these traits could be plastic to environmental conditions, as are other morphological and functional traits. Given the ecological significance of dispersal strategies, fruit characteristics, and their plasticities in plant invasiveness, these characteristics should be evaluated more extensively (Doudová et al., 2017).
Bidens bipinnata is native to East Asia (Wang et al., 2017). B. frondosa, which originated in North America, and B. pilosa, originated in tropical America, have been reported in the Korean Peninsula since 1964 (Park et al., 2002). B. frondosa occurs throughout South Korea, but B. pilosa is mainly found near the coastal area, especially around major harbors of South Korea (Figure 1) (Korea National Arboretum, 2016). We treated B. pilosa as a noninvasive alien in this study because its distribution is restricted to the candidate origin of introduction despite the maintenance of natural populations (Blackburn et al., 2011;Mathakutha et al., 2019). Achenes of Bidens species have pappi that consist of two to three barbed awns that facilitate adhesive dispersal (Sorensen, 1986). Since plant species with smaller seed sizes or longer awns tend to have a higher probability of remaining attached to animal fur (Ansong & Pickering, 2014;Kiviniemi & Telenius, 1998), both achene size and pappus length in Bidens species likely affect dispersal efficiency (Rocha, 1996).
Phenotypic plasticity is defined as the differential phenotypic expression of a genotype in response to environmental factors, so an experimental study using clonal replicates is ideal for examining phenotypic plasticity (Pigliucci, 2001). However, ecological studies often evaluate plasticity at the population or species level by comparing the average trait value of a population or species between environmental treatments (Davidson et al., 2011;Richards et al., 2006).
Although the results might confound plasticity with genotypic variation between treatments, evaluation of phenotypic plasticity at the species level has an advantage of including diverse genotypes representing a species. If there is no bias in assigning individual plants to environmental treatments, plasticity at the species level can provide information on the ecological significance of plasticity in invasive species (Davidson et al., 2011;Richards et al., 2006).
Here, we examined the phenotypic expression of native and alien Bidens species in response to nutrient availability and disturbance.
Specifically, the following questions were addressed: (a) Do native, noninvasive, and invasive alien Bidens species exhibit differential morphological and functional traits? (b) Do Bidens species respond differently to fertilization and competition treatments? (c) Does invasive Bidens species have morphological characteristics that better facilitate long-distance dispersal compared to noninvasive species?

| Experimental design
Achenes of Bidens species were sterilized using 0.5% sodium hypochlorite for 10 min, carefully scarified twice with a razor blade, and maintained in Petri dishes with wet filter paper in June 2016.
After cold treatment at 4°C for 2 weeks, achenes were allowed to germinate in a customized walk-in chamber under a 12-hr light/ dark photoperiod at 23°C. Individual seedlings were planted in flats containing commercial soil medium (ShinSung Mineral Co. Ltd.) and grown under the same conditions present during germination. To evaluate the effects of competition and fertilization, we employed a full factorial design with four treatment combinations.
As a competition treatment, we transplanted one individual seedling of each Bidens species to the center of pots containing Festuca arundinacea (tall fescue). F. arundinacea was chosen as a competing species because it has been widely used to recover anthropogenically disturbed areas in Korea. Before transplantation, 1 g of F. arundinacea seeds was sowed in each pot following the distributor's recommendations and grown for one and a half months in the common garden. To simulate the high-nutrient content of soil in agricultural areas, half of the pots in each competition treatment

| Trait measurement
After 6 weeks of treatment, we collected whole plants, washed their roots to remove all soil, and dried them in a dry oven (Hanbaek, Co. Ltd.) at 65°C for three days to weigh their dry mass.
The R/S ratio was calculated as the dry mass of the root divided by the aboveground biomass. Before harvesting, we took one fully expanded leaf from each plant and scanned it to measure the leaf area (Digimizer software ver. 4.6.1; MedCalc Software bvda).
Leaves were dried at 65°C for 3 days, and the SLA was calculated as the leaf area divided by the dry mass. We estimated the leaf chlorophyll content using a SPAD-520 plus chlorophyll meter (Spectrum Technologies). The SPAD values exponentially correlate with the chlorophyll content.
We counted the number of inflorescences as a proxy of fecundity, because many fruits fell onto the ground during the experiment.
Bidens species produce morphologically distinctive achenes at the center and periphery of the capitulum (Brändel, 2004). Disk achenes tend to be larger and disperse more readily than ray achenes (Rocha, 1996). We randomly selected more than two achenes from the central area of the capitulum of each plant individual and photographed them using a stereomicroscope. Images were analyzed using Digimizer software to measure the length of the pappi and the cross-sectional areas of the achenes. Achenes with broken pappi were excluded from the image analysis, and morphological characteristics were analyzed for a total of 250 B. bipinnata achenes, 290 B. frondosa achenes, and 140 B. Pilosa achenes.

| Statistical analyses
In order to evaluate the phenotypic plasticity of Bidens species, we computed the relative distance plasticity index (RDPI) of each trait following Valladares et al. (2006). For each species, we calculated relative distances of trait values between all pairs of individuals that were grown in different environments. The RDPI for the competition was the average of the relative distances between competition treatments in the absence and presence of fertilization. Similarly, the RDPI for the fertilization was the average of the relative distances between fertilization treatments in the absence and presence of the competition.
All statistical analyses were performed using the R statistical package version 3.2.4 (R Foundation for Statistical Computing). To compare morphological and functional traits among treatments and species, three-way analyses of variance (ANOVA) were conducted using aov function. The measured traits were the dependent variable, and the fertilization treatment, competition treatment, species, and their interactions were independent variables. In order to interpret the species by treatment interactions, we conducted two additional analyses. First, to examine whether testing species exhibited differential trait values in each treatment, differences between species were evaluated for each treatment with post hoc Tukey mean comparison test (tukeyHSD function). To assess differential treatment effects among species, we conducted two-way ANOVA for each species with environmental treatments and their interactions as independent variables. The total biomass, shoot biomass, and cross-sectional area of achenes were log-transformed to meet the normality assumption. To examine whether Bidens species showed differential phenotypic plasticity, one-way ANOVA was conducted with the relative distances of a trait as dependent variable and species as independent variable. Differences between species were evaluated using post hoc Tukey mean comparison test.  Abbreviations: Comp, competition; Fert, fertilization; R/S ratio, root-to-shoot ratio; SLA, specific leaf area.

| Vegetative and functional traits
When averaged across treatments, alien invasive B. frondosa and noninvasive B. pilosa produced more biomass than did native B. bipinnata (Table 1, Figure 2). All tested species increased their total biomass and shoot biomass in response to the fertilization treatment. In contrast, species responded to the competition treatment differently, as indicated by significant species by competition interactions ( than those of the other species (Figure 2d).

F I G U R E 2
Effects of environmental treatments on the morphological and functional traits of three Bidens species. See Table 1 and Table S1 for the significance tests. Letters indicate statistically significant differences among species and treatments at the 0.05 level based on Tukey's adjustment. R/S ratio, root-to-shoot ratio, SLA, specific leaf area Responses of other functional traits also differed among the in-  Figure 2g). B. frondosa exhibited a lower R/S ratio than did the other species.
In response to environmental treatments, testing Bidens species showed differential RDPIs in all measured vegetative and functional traits except the shoot biomass responding to the fertilization treatment ( Figure 3). RDPI of B. frondosa was higher than those of noninvasive alien B. pilosa (Figure 3). B. frondosa exhibit higher RDPI than did native B. bipinnata in the total biomass, root biomass, and final height (Figure 3a,c,d), but RDPI of those two species was similar in the SLA and chlorophyll contents (Figure 3e,f).

| Reproductive and achene traits
All tested species produced more inflorescences in response to the fertilization treatment. In contrast, the effect of competition on the inflorescence number differed among the tested species, as indicated by the statistically significant species by competition interaction (  Figure 2i,j). The achene morphology of the tested species responded to the fertilization treatment differently, as indicated by the significant species by fertilization interactions (Table 1).
Under fertilization treatment, the pappus length of the achenes de-

| Vegetative and functional traits
Massive biomass, tall stature at maturity, large SLA, and high chlorophyll content have been proposed as major attributes of invasive plant species (Divišek et al., 2018;Gallagher et al., 2014;van Kleunen et al., 2010;Leishman et al., 2007). These traits of testing Bidens species responded to the fertilization and competition treatments, but the responses differed among plant species (Figure 2). In particular, traits of noninvasive alien B. pilosa barely responded to the environmental treatments. Invasive B. frondosa exhibited higher RDPI than did B. pilosa (Figure 3). Even though environmental treatments affected the trait expression of both invasive B. frondosa and native B. bipinnata, B. frondosa had higher RDPI than those of native B. bipinnata did in the total biomass and final height (Figure 3). These results support the hypothesis that invasive plants tend to be more plastic than congenic native and noninvasive alien plants (Davidson et al., 2011).
Due to the higher plasticity, B. frondosa had more massive biomass after six weeks of growth and a greater final height than native B. bipinnata. Thus, the higher plasticity of B. frondosa regarding biomass and height likely confers a competitive advantage over native B. bipinnata (van Kleunen et al., 2015). Notably, differences in the The representative functional traits of invasive plant species include high SLAs and low R/S ratios, which are associated with greater resource capture abilities and faster growth rates (Gallagher et al., 2014;van Kleunen, Weber, et al., 2010;Leishman et al., 2007).
The R/S ratios of the three investigated species were not plastic in response to environmental treatments, and invasive B. frondosa exhibited lower R/S ration than did native and alien noninvasive Bidens species across treatments. The homeostatic maintenance of a low R/S ratio would be beneficial particularly in a high-nutrient environment (Gioria & Osborne, 2014); this likely explains the plastic increase of shoot biomass observed in B. frondosa. The three investigated species exhibited similar SLAs in the no-competition treatment, while the SLA of B. frondosa was larger than that of the other species in the competition treatment. Since the effects of high SLA on the resource capture ability would be manifest in a high-nutrient environment without competition (Gioria & Osborne, 2014), the contribution of SLA to the competitive ability of B. frondosa would be limited.
In summary, we could not find evidence that the plasticity of vegetative and functional traits contributes to invasive success.
Instead, the absolute trait values rather than the plasticity of the R/S ratios likely contribute to invasive success (Godoy et al., 2012; Matzek, 2012).

| Reproductive traits
Invasive B. frondosa produced more inflorescences than did native and noninvasive alien Bidens species, though this difference was manifested only when plants were grown without competition. Such a fecundity advantage would enhance the invasive success of B. frondosa and provide a competitive advantage over native B. bipinnata.
B. pilosa and B. bipinnata exhibited a similar number of inflorescences across environmental treatments.
A previous meta-analysis showed that the fitness components of invasive plants tended to be more plastic than those of native and noninvasive alien plants in response to nutrient availability (Davidson et al., 2011). Notably, this conclusion has been challenged, as the selection of native species for comparison might affect the results. For instance, alien and native weedy species exhibited similar responses to fertilization and competition (Dawson et al., 2012).
Our results do support the hypothesis associating higher plasticity of fitness components with invasive species.
In this study, the total biomass and inflorescence number exhibited similar plasticity patterns in response to the experimental treatments. However, only the inflorescence number showed differential trait values among the tested species. The total biomass of B. frondosa and B. pilosa was similar even though the plasticity of B. frondosa was higher than that of B. pilosa. While fecundity is a more appropriate fitness measure than biomass for annual plant species (Crawley, 1997), total biomass has been widely used as a proxy of fitness based on the observation that biomass is highly correlated with reproductive output (Davidson et al., 2011;Dawson et al., 2012;Weiner et al., 2009). Our results suggested that the choice of fitness components might lead to contrasting conclusions, and diverse fitness components should be examined to evaluate the significance of traits associated with invasive success.
Since disturbance increases resource availability (Jauni et al., 2015), we expected that the fertilization and no-competition treatments would likely affect the performance of the investigated plants in a similar manner. However, distinct effects were observed under these treatments; fertilization increased the inflorescence number of B. pilosa, but this effect was not detected under the no-competition treatment (Figure 2h). In B. frondosa, the plastic inflorescence-number response to the no-competition treatment was higher than the response to the fertilization treatment.
Therefore, the effects of disturbance may not be straightforward, as the removal of neighboring plants via disturbance would influence not only resource availability but also other biological interactions, such as the physical interference of root growth patterns (Padilla et al., 2013).
Invasive plants tend to have smaller seed sizes, potentially increasing their dispersal distance (Rejmánek & Richardson, 1996). In Bidens species, however, the length of pappi, in addition to the achene size, should be considered to evaluate the dispersal potential, since pappus length would affect the attachment and retention of achenes to animal furs (Couvreur et al., 2004;Kiviniemi & Telenius, 1998;Sorensen, 1986). Studies comparing seeds from multiple species found that seeds with more extensive spines were transported longer than seeds with short awns (Ansong & Pickering, 2016;Couvreur et al., 2004), indicating that a longer seed appendage likely has a positive effect on seed dispersal. In this study, the achenes of B. pilosa were smaller and featured longer pappi than those of B. frondosa (Figure 2i,j). Given that dispersal distance via adhesion has a positive relationship with the length of seed appendages but a negative correlation with the achene size, the noninvasive alien B. pilosa had a higher dispersal potential than did invasive alien B. frondosa. The achenes of B. bipinnata had the longest pappi and largest achene size among the three tested species, so it remains unclear whether it has a lower dispersal potential than the other species.
Long-distance dispersal has been suggested as a key characteristic of the rapid range expansion of invasive species (Coutts et al., 2011;Rejmánek & Richardson, 1996). However, it should be noted that the fitness advantage of more distant dispersal depends on the ecological context. For instance, seeds that disperse further from maternal plants likely have a higher probability of facing novel environmental conditions. Because annual and biennial plants die after reproduction, no competition between maternal plants and offspring occurs. Thus, if the habitats of maternal plants provide favorable conditions for plant growth, annual plants would reap greater benefits from a short dispersal distance than from distant dispersal to novel environmental conditions. Through short-distance dispersal, a population possibly accumulates large numbers of fruits that are sufficient to disperse into a novel habitat and establish there (Coutts et al., 2011;Hemrová et al., 2017).
Since B. frondosa mainly occurs in agricultural fields with nutrient-rich soil, a shorter dispersal distance and the accumulation of propagules would better promote its rapid expansion than would long-distance dispersal.
The pappus length and achene size of the investigated species responded to nutrient addition differently. The pappus length of native B. bipinnata increased under the fertilization treatment, which likely increased its dispersal potential. In contrast, B. frondosa had shorter pappi and larger achene areas when grown in nutrient-rich soil; therefore, its dispersal ability is expected to decrease under such conditions. The plasticity of the achene morphology in B. frondosa likely allows it to occupy favorable habitats, thereby increasing its population size and expansion rate.

| CON CLUS IONS
Our results demonstrated that the invasive success of B. frondosa is likely attributable to its higher plasticity regarding the number of in-