Functional traits contributed to the superior performance of the exotic species Robinia pseudoacacia: a comparison with the native tree Sophora japonica

Abstract

Functional traits determine the ecological strategies of plants and therefore are widely considered to feature in the success of invasive species. By comparing a widespread exotic invasive species Robinia pseudoacacia L. with a related native one Sophora japonica L., this research aimed to study strategies of R. pseudoacacia for superior performance from the perspective of functional traits. We conducted a greenhouse experiment in which seedlings of R. pseudoacacia and S. japonica were grown separately under a factorial combination of two light regimes and three levels of nitrogen (N) fertilization, including a control and two levels intended to represent ambient and future levels of N deposition in Chinese forests. After 90 days of treatment, performance and functional traits were determined for the two species, the former referred to as the total biomass (TB) that directly affected fitness. Trait plasticity and integration (the pattern and extent of functional covariance among different plant traits) were analyzed and compared. We found that the two species showed significantly different plastic responses to light increase: in the low-light regime, they were similar in performance and functional traits, while in the high-light regime, R. pseudoacacia achieved a significantly higher TB and a suite of divergent but advantageous functional traits versus S. japonica, such as significantly greater photosynthetic capacity and leaf N concentration, and lower carbon-to-N ratio and root-to-shoot ratio, which conferred it the greater performance. Moreover, across the light gradient, R. pseudoacacia showed higher correlations between photosynthetic capacity and other functional traits than S. japonica. In contrast, N deposition showed little impact on our experiment. Our results suggested that across light regimes, three aspects of functional traits contributed to the superior performance of R. pseudoacacia: functional trait divergence, significantly different plasticity of these traits, as well as greater overall trait coordination.

Keywords: exotic plant invasion, functional trait divergence, light gradient, phenotypic plasticity, species comparison, trait coordination.

Introduction

Exotic plant invasion has aroused widespread concern due to its enormous negative impacts on native biodiversity and ecosystem functioning (Sala et al. 2000, Zhang et al. 2009, Vilà et al. 2011). Since functional traits determine the responses of plants to complex environments (Suding et al. 2008), it is widely accepted that the invasion of exotic plants in novel environments largely relies on their functional traits (Richardson and Pyšek 2006, Tecco et al. 2010). Based on comparisons with native species, three aspects have been proposed to contribute to the superior performance of invasive species: trait values, phenotypic plasticity and phenotypic integration (Davidson et al. 2011, Godoy et al. 2012, Matzek 2012).

To explain how functional trait values of invasive species relative to coexisting natives promote invasion success, two contrasting hypotheses have been put forward. The first one is phenotypic convergence (Thompson et al. 1995, Duncan and Williams 2002): mainly based on the process of habitat filtering, this hypothesis emphasizes the importance of environmental factors and maintains that invasive and coexisting native species show strong similarity in functional traits. Therefore, exotic species can adapt to local environments and invade native communities more easily (Duncan and Williams 2002, Daehler 2003). This hypothesis has been favored by some empirical studies, such as Scharfy et al. (2011) and Drenovsky et al. (2012). Alternatively, mainly based on the theory of leaf economic spectrum and limiting similarity (Abrams 1983, Wright et al. 2004), the phenotypic divergence hypothesis holds that the larger the difference in functional traits between invasive and native species, the greater the probability for invasive species to succeed (Ordonez et al. 2010). Many studies have suggested functional trait divergence (Penuelas et al. 2010, Godoy et al. 2011, Heberling and Fridley 2013). Meta-analyses and large-scale studies indicate that invasive species generally possess high specific leaf area (SLA), net photosynthetic rate and tissue nutrient concentration, and low root-to-shoot ratio (RSR), which result in a faster return on investments and a higher growth rate (Leishman et al. 2007, van Kleunen et al. 2010). Despite large amounts of research, universal conclusions have not been reached.

Phenotypic plasticity, defined as the ability of a certain genotype to alter its phenotypic expression in reaction to changing environments (Schlichting 1986, Davidson et al. 2011), is also considered to feature in the invasion success of exotic species. Richards et al. (2006) put forward that the plasticity of functional traits may allow invasive species to better acclimate to fluctuating environmental conditions. Based on comparisons with native species, three primary scenarios have been proposed: (i) Jack-of-all-trades, through functional trait plasticity, invasive species are better able to maintain fitness under unfavorable conditions than natives; (ii) master-of-some, invasive species can better take advantage of increased resources and increase fitness to a larger extent than natives; and (iii) Jack-and-master, invasive species are able to combine both the above characteristics. Theoretically, higher phenotypic plasticity may broaden the ecological niche breadth of invasive species and help them to better adapt to and establish in a wide range of environments. Numerous studies have demonstrated that invasive species exhibit greater trait plasticity relative to coexisting natives (Funk 2008, Davidson et al. 2011, Paquette et al. 2012), yet there is also a mass of evidence supporting similar levels of plasticity (Palacio-López and Gianoli 2011, Scharfy et al. 2011, Kuehne et al. 2014). This inconsistency may arise from the fact that trait plasticity does not necessarily confer a fitness advantage (Richards et al. 2006).

Higher phenotypic integration may also contribute to the success of invasive species (Godoy et al. 2012, Osunkoya et al. 2014). Referring to the pattern and extent of functional covariance among different plant traits, phenotypic integration is usually estimated as the number of significant correlated trait pairs (Pigliucci 2003, Gianoli and Palacio-López 2009). Because of the coordinated variation among traits, an integrated phenotype may better exploit the environment, cope with environmental heterogeneity more efficiently and reduce the cost of non-adaptive and maladaptive plasticity (van Kleunen and Fischer 2005, Osunkoya et al. 2010). Thus, phenotypic integration is considered to be positively related to fitness and environmental stress (Gianoli 2004). However, experimental research also found that phenotypic integration may restrict the expression of phenotypic plasticity, which was an unexpected result since the two aspects were both considered to promote invasion (Gianoli and Palacio-López 2009). Our understanding of phenotypic integration is still insufficient and more research is required (Godoy et al. 2012). Besides, recent literature reviews show that most work exploring how functional traits enhance the performance of invasive species only focuses on trait values and plasticity (Ordonez et al. 2010, Godoy et al. 2011, Matzek 2012).

Robinia pseudoacacia L., native to North America, is now a widespread exotic species across China and is usually dominant in warm temperate forests (Wang and Zhou 2000). Although listed as one of the predominant invaders in the world (Boring and Swank 1984, Cierjacks et al. 2013), it is still widely used for reforestation in north China. In order to reveal how functional traits contribute to the superior growth performance of R. pseudoacacia, the native species Sophora japonica L. was chosen for comparison. Sophora japonica is widely distributed throughout China but rarely dominant (Wang and Zhou 2000); both species belong to the same subfamily (Papilionoideae). With a greenhouse experiment, we compared the performance of the above two species across environmental gradients and explored their functional trait values, phenotypic plasticity and integration. Light and nitrogen (N) deposition were chosen as environmental factors. In forest ecosystems, they are both under fluctuation in the context of human disturbance and global change, which may play important roles in plant growth (Guo et al. 2006, Reay et al. 2008). Specifically, we addressed the following questions:

1. How does R. pseudoacacia compare with S. japonica in performance (evaluated by total biomass (TB)) under different treatments? Does R. pseudoacacia show convergence or divergence in functional traits relative to S. japonica?

2. Is R. pseudoacacia more plastic in traits than S. japonica in response to different treatments?

3. Does R. pseudoacacia show higher phenotypic integration than S. japonica across environmental gradients?

Materials and methods

Study site

This study was conducted at the Fanggan Research Station of Shandong University (36°26′N, 117°27′E), which is located in the Central Mountainous Region of Shandong Province, China. Characterized by a warm temperate monsoon climate, this area has a mean annual temperature of 13 ± 1 °C and a mean annual precipitation of 700 ± 100 mm. The main vegetation type is mixed forests of the warm temperate zone (Zhang et al. 2006). To keep a homogenous environment, the whole experiment was performed in a greenhouse made up of a steel pipe frame, which was covered by a plastic film. Throughout the experiment, the greenhouse was well ventilated by rolling up the side film in the daytime. During the growth period, the microclimate in the greenhouse was monitored (MicroLog EC650, Fourtec, Contoocook, NH, USA): in the daytime the average temperature and relative humidity varied from 21.1 ± 0.2 to 38.3 ± 2.4 °C and 27.9 ± 3.1 to 87.3 ± 0.6%; in the night they ranged from 12.1 ± 0.9 to 27.7 ± 0.3 °C and 46.0 ± 2.5 to 94.5 ± 0.5%.

Plant materials

Seeds of R. pseudoacacia and S. japonica were bought from Dacheng Seed Company (Jinan, China), which gathered them in its own garden in late autumn of 2012. In late April of 2013, after soaking in distilled water for 24 h, the seeds were allowed to germinate in a growth chamber. When most radicles reached 2 cm, healthy and uniform germinated seeds were sown in plastic pots (25 × 21 cm, height × diameter) with 6 kg loam and 2 kg sand, which were carefully sieved and thoroughly mixed. The chemical properties of the substrates were determined as follows: pH 6.51, available N 50.20 mg kg⁻¹ and available phosphorus (P) 31.14 mg kg⁻¹.

Experimental design

We designed a factorial experiment for each species. Seedlings were exposed to two light regimes. The high-light regime (66% of full radiation), corresponding to the light conditions in canopy gaps of deciduous broad-leaved forests in this area, was achieved with the plastic film of the greenhouse; while the low-light regime (10% of full radiation), representing the understorey light conditions, was created by a black nylon net shelter in the greenhouse (Du 2010). In each light regime, three N deposition levels were applied: 0, 3 and 9 g N m⁻² year⁻¹. The 0 g N m⁻² year⁻¹ level was set as control; 3 and 9 g N m⁻² year⁻¹ represented a moderate deposition level already reported in China and a higher level probably reached in the future, respectively. Nitrogen deposition was simulated through adding mixed solutions of (NH₄)₂SO₄ and KNO₃ (1 : 1, M/M), considering that in recent years the ratio of NH₄-N and NO₃-N in atmospheric N deposition was about 2 in China (Liu et al. 2013). Different concentrations of K₂SO₄ and KCl solutions were also added accordingly to make sure that all treatments received the same quantity of potassium as well as sulfur. For each species, there were six treatment combinations in total and six pots were randomly assigned to each combination as replicates.

All treatments started from 1 June 2013, when the seedlings were 1 month old, and ended on 30 August 2013. The N deposition was applied biweekly, seven times in total. According to the report that in north China, ∼70% of the annual precipitation falls between June and August (Wang and Zhou 2000), solutions equivalent to 10% of the annual N deposition were applied each time (for compositions and concentrations of solutions see Table S1 available as Supplementary Data at Tree Physiology Online). During the experiment, all the seedlings were well watered. Weeds and insects were removed manually. Pots in each light regime were rearranged randomly during the experiment.

Measurement of plant traits

In this experiment, two sets of traits were measured: the performance trait and functional traits. The former referred to TB that directly influenced fitness, while the latter were those that contributed to fitness indirectly through their impacts on growth (Violle et al. 2007, Ruprecht et al. 2014). Functional traits were divided into two categories, morphological and physiological traits, as shown in Table 1.

Table 1.

Results of three-way ANOVA for the effects of species identity (I), light (L), nitrogen deposition (N) and their interactions on performance and functional traits of R. pseudoacacia and S. japonica. TB, total biomass (g); RSR, root-to-shoot ratio; SLA, specific leaf area (cm² g⁻¹); LAR, leaf area ratio (cm² g⁻¹); A_max, maximum net photosynthetic rate (μmol CO₂ m⁻² s⁻¹); N_mass, leaf nitrogen concentration (mass based) (mg g⁻¹); C : N, leaf C-to-N ratio; Chl_mass, leaf chlorophyll concentration (mass-based) (mg g⁻¹); Chl a : b, leaf chlorophyll a-to-chlorophyll b ratio; Chl : N, leaf chlorophyll-to-leaf nitrogen ratio; WUE, instantaneous water-use efficiency (mmol mol⁻¹). TB, RSR, N_mass, C : N: n = 4–5; A_max, WUE: n = 3–4; SLA, LAR, Chl_mass, Chl a : b, Chl : N: n = 3. Numbers in the table represent F-ratios; F-ratio subscripts represent numerator and denominator degrees of freedom for each effect; overall trait means show the mean ± SE of traits across the treatments; asterisks indicate significant effects: ***P ≤ 0.001, **P ≤ 0.01 and *P ≤ 0.05; significantly different trait means of R. pseudoacacia and S. japonica are denoted by bold font.

	I	L	N	I × L	I × N	L × N	I × L × N	Overall trait means
								R. pseudoacacia	S. japonica
Performance trait
TB	16.911,47***	58.291,47***	4.432,47*	4.091,47*	0.762,47	2.062,47	0.972,47	50.00 ± 5.88	30.24 ± 3.11
Morphological traits
RSR	8.331,47**	56.921,47***	3.002,47	19.871,47***	3.542,47*	1.542,47	2.402,47	0.378 ± 0.023	0.499 ± 0.064
SLA	2.091,24	475.941,24***	0.022,24	2.621,24	4.682,24*	0.862,24	0.952,24	399.27 ± 46.01	376.66 ± 39.98
LAR	0.051,24	343.701,24***	0.362,24	0.011,24	0.392,24	0.192,24	0.062,24	147.64 ± 20.60	149.69 ± 21.10
Physiological traits
Amax	8.781,25**	66.591,25***	1.722,25	28.251,25***	0.602,25	0.602,25	0.492,25	11.43 ± 1.09	9.62 ± 0.48
Nmass	32.681,47***	163.151,47***	6.992,47**	6.411,47*	2.922,47	4.592,47*	4.202,47*	28.03 ± 1.23	23.21 ± 1.39
C : N	45.881,47***	225.231,47***	10.762,47***	29.371,47***	1.692,47	8.772,47***	3.452,47*	16.05 ± 0.82	20.05 ± 1.30
Chlmass	11.491,24**	62.211,24***	0.022,24	0.401,24	0.012,24	0.012,24	0.272,24	4.14 ± 0.35	3.20 ± 0.27
Chl a : b	13.571,24***	6.201,24*	0.542,24	0.361,24	0.182,24	1.962,24	1.482,24	1.71 ± 0.04	1.56 ± 0.01
Chl : N	1.011,24	5.301,24*	0.552,24	0.851,24	0.642,24	0.092,24	0.222,24	0.153 ± 0.009	0.142 ± 0.006
WUE	2.951,25	1.211,25	0.092,25	0.861,25	0.232,25	0.382,25	0.192,25	3.32 ± 0.15	3.96 ± 0.30

After 65 days of treatment, using a portable leaf gas exchange system (GFS-3000, WalzGmbH, Effeltrich, Germany), leaf gas exchange parameters were measured between 8:30 and 12:00 h (local time) on two consecutive sunny days. For each species, three to four fully expanded and healthy leaves from the upper shoots (one leaf per seedling) were selected from each treatment. The maximum net photosynthetic rate (A_max) and transpiration rate (E) were measured under a photosynthetic photon flux density of 1000 mol m⁻² s⁻¹ (a saturated light intensity derived from a pre-experiment, data not shown) provided by a red-blue light emitting diode. The instantaneous water-use efficiency (WUE) was calculated as A_max/E. During the measurement, the air temperature, relative humidity and CO₂ concentration in the leaf cuvette were set as 27.5 °C, 60% and 400 mol mol⁻¹, respectively.

Leaf chlorophyll parameters were measured in the middle of August. For each species, three fully expanded and healthy leaves from the upper shoots (one leaf per seedling) in each treatment were sampled and measured for fresh weight. After extraction by 95% ethanol (v/v), the concentrations of leaf chlorophyll a and b and total chlorophyll (Chl_mass, mass based) were measured according to the spectrophotometric method of Lichtenthaler and Wellburn (1983). The ratio of leaf chlorophyll a and b (Chl a : b) was also calculated.

After 88 days of treatment, leaf morphology was measured on 28 August, which was the 119th day after germination. For each species, three fully expanded and healthy leaves from the upper shoots (one leaf per seedling) in each treatment were scanned and images were analyzed with the WinFOLIA Pro 2009a software (Regent Instruments, Inc., Quebec, QC, Canada) to obtain leaf areas. The leaves were then oven-dried at 80 °C for 48 h and weighed. Specific leaf area (SLA) was calculated as leaf area/leaf dry weight.

At the end of the experiment, four to five seedlings in each treatment were harvested for each species. The whole plant was divided into three parts: root, stem and leaf. Roots were washed carefully with water to remove soil and debris. All parts were oven-dried at 80 °C for 48 h and weighed. Total biomass and biomass allocation parameters were calculated as follows:

$$\begin{array}{ll}& \text{Total biomass (TB)}\\ & =\text{root biomass}+\text{stem biomass}+\text{leaf biomass;}\\ & \text{Root-to-shoot ratio (RSR)}=\frac{\text{root biomass}}{(\text{stem biomass}+\text{leaf biomass)}};\\ & \text{Leaf area ratio (LAR)}=\text{SLA}\times \frac{\text{leaf biomass}}{\text{total biomass}}.\end{array}$$

After the determination of biomass parameters, all the leaf samples were kept for the measurement of leaf N and carbon (C) concentrations (N_mass and C_mass, mass-based) with the Kjeldahl method and the potassium dichromate volumetric method, respectively. Leaf C-to-N ratio (C : N) and leaf chlorophyll-to-N ratio (Chl : N) were calculated accordingly. The elemental analyses of leaves were completed at Shandong Agricultural University.

Statistical analyses

To test the main effects and interactions of species identity, light and N deposition, three-way analysis of variance (ANOVA) was performed on plant performance and functional traits. Data were log-transformed when necessary to meet the normality and homogeneity of variance assumptions of ANOVA. Compared with species identity and light, the N effects and the interactions between species identity and N were minimal (see Results). Therefore, Tukey's HSD test was only conducted for significant species identity × light interactions to compare the two species across different light regimes. When there was no significant interaction between species identity and light, independent sample t-tests were performed between the two species under each light regime separately to distinguish them. All the above analyses were accomplished with the SPSS 16.0 software (SPSS Inc., Chicago, IL, USA).

For each species, we used the reaction norm of trait to evaluate phenotypic plasticity: significant species identity × light interactions were interpreted as species differences in trait plasticity in response to light (Mozdzer and Megonigal 2012, Chen et al. 2013). Defined as the array of trait values produced under different environments by a single genotype (Pigliucci and Schlichting 1995), the reaction norm is considered as the most classic and immediate approach to estimate phenotypic plasticity.

Lastly, to explore how the two species were separated in a multivariate context and how plant traits were influenced by experimental treatments, a principal component analysis (PCA) was conducted on the correlation matrix of the standardized data of all functional traits. Ordinary least-square regressions were carried out on plant TB and the first two PCA axes for the two species separately. To evaluate the level of phenotypic integration for each species, Pearson product–moment correlation coefficients were generated among all the functional traits. Principal component analysis was implemented with the stats package in R 2.15.3 (R Development Core Team 2013); Pearson correlation was performed with the SPSS 16.0 software (SPSS Inc.).

Results

General effects and trait values

Species identity and light had much greater effects than N deposition (Table 1), which only significantly affected TB, N_mass and C : N. Since species identity and N did not show significant interactions on the three parameters, one-way ANOVAs were performed for each species separately to examine the effects of N deposition. Results showed that for R. pseudoacacia, N did not significantly affect TB (F_{2, 27} = 0.754, P = 0.480), N_mass (F_{2, 26} = 2.147, P = 0.136) and C : N (F_{2, 26} = 2.367, P = 0.113), and this was also the case for S. japonica (TB: F_{2, 27} = 2.257, P = 0.125; N_mass: F_{2, 27} = 0.598, P = 0.557; C : N: F_{2, 27} = 0.702, P = 0.505). Therefore, in this experiment, simulated N deposition showed little impact on either species and we mainly analyzed the effects of species identity and light.

As indicated by the significant effect of species identity on TB (Table 1), in general, R. pseudoacacia showed higher performance than S. japonica. However, TB of R. pseudoacacia was only significantly greater in the high-light regime; in the low-light regime, no difference was detected between the two species (Figure 1). In terms of functional traits, on the whole, R. pseudoacacia had significantly higher A_max, N_mass, Chl_mass and Chl a : b than S. japonica, while its RSR and C : N were significantly lower (Table 1). Similar to TB, RSR, A_max, N_mass and C : N of the two species were only significantly different in the high-light regime; in the low-light regime, values of these traits were similar (Figures 2a and 3a–c).

Figure 1.

Reaction norms of TB in response to light regimes for R. pseudoacacia and S. japonica. The values are shown as mean ± SE (n = 14–15). Different letters denote significant differences (P ≤ 0.05) with Tukey's test. LL, low light; HL, high light.

Figure 2.

Reaction norms of morphological traits in response to light regimes for R. pseudoacacia and S. japonica. RSR, n = 14–15; SLA, LAR, n = 9. See Table 1 title for trait abbreviations. The values are shown as mean ± SE. Different letters denote significant differences (P ≤ 0.05) with Tukey's test. LL, low light; HL, high light.

Figure 3.

Reaction norms of physiological traits in response to light regimes for R. pseudoacacia and S. japonica. A_max, WUE, n = 9–10; N_mass, C : N, n = 14–15; Chl_mass, Chl a : b, Chl : N, n = 9. See Table 1 title for trait abbreviations. The values are shown as mean ± SE. Different letters denote significant differences (P ≤ 0.05) with Tukey's test. Asterisks indicate significant differences between species in each light regime separately with independent sample t-test: ***P ≤ 0.001, **P ≤ 0.01, *P ≤ 0.05. LL, low light; HL, high light.

Trait plasticity

Species identity × light interactions were significant for many of the traits examined (TB, A_max, RSR, N_mass and C : N), indicating that the two species differed in their reaction norms and thus plasticity to light. In contrast, the identity × N, light × N and identity × light × N interaction effects were minimal, suggesting the consistent responses of the two species to N under different light conditions (Table 1). Both species responded to increased light with enhanced biomass production and photosynthesis, but R. pseudoacacia showed significantly greater responses, which indicated higher phenotypic plasticity for TB and A_max (Figures 1 and 3a). On the contrary, RSR, N_mass and C : N of R. pseudoacacia exhibited smaller reactions to light relative to S. japonica, suggesting lower plasticity for these traits (Figures 2a and 3b and c).

Trait correlation

For each species, 45 pairs of functional traits were examined for correlation. Among these traits, 30 pairs for R. pseudoacacia and 26 pairs for S. japonica were significantly correlated at P ≤ 0.05 (Table 2). Obviously, the overall strength of trait correlation was similar between the two species. However, an intriguing phenomenon was observed: for R. pseudoacacia, A_max was significantly correlated with all the other functional traits measured, while for S. japonica, it was only significantly correlated with Chl a : b.

Table 2.

Matrix of Pearson product–moment correlation coefficients for functional traits of R. pseudoacacia and S. japonica across different light regimes. See Table 1 title for trait abbreviations. Significant correlations are denoted by bold font and asterisks: ***P ≤ 0.001, **P ≤ 0.01 and *P ≤ 0.05.

Traits	Amax	Nmass	C : N	Chlmass	Chl a : b	Chl : N	WUE	RSR	SLA	LAR
R. pseudoacacia (n = 18)
Amax	1.000
Nmass	−0.643**	1.000
C : N	0.587*	−0.974***	1.000
Chlmass	−0.761***	0.736***	−0.716***	1.000
Chl a : b	0.537*	−0.141	0.105	−0.407	1.000
Chl : N	0.827***	−0.125	0.040	−0.454	0.584*	1.000
WUE	0.562*	−0.501*	0.352	−0.343	−0.012	0.445	1.000
RSR	0.683**	−0.451	0.372	−0.639**	0.475*	0.587*	0.254	1.000
SLA	−0.938***	0.782***	−0.740***	0.815***	−0.356	−0.663**	−0.600**	−0.631**	1.000
LAR	−0.927***	0.797***	−0.736***	0.757***	−0.379	−0.648**	−0.632**	−0.669**	0.974***	1.000
S. japonica (n = 18)
Amax	1.000
Nmass	−0.279	1.000
C : N	0.301	−0.975***	1.000
Chlmass	−0.149	0.869***	−0.902***	1.000
Chl a : b	0.542*	−0.573*	0.516*	−0.413	1.000
Chl : N	0.277	0.818***	−0.802***	0.779***	−0.290	1.000
WUE	0.046	−0.017	0.181	−0.170	−0.155	−0.078	1.000
RSR	0.015	−0.768***	0.816***	−0.764***	0.171	−0.727***	0.446	1.000
SLA	−0.392	0.885***	−0.914***	0.890***	−0.551*	0.654**	−0.252	−0.745***	1.000
LAR	−0.346	0.900***	−0.913***	0.898***	−0.576*	0.705**	−0.227	−0.739***	0.985***	1.000

Principal component analysis on functional traits of the two species showed that the first three axes explained 87.7% of the total variation. The first axis (with 53.2% of data variation captured) was strongly associated with RSR, SLA, LAR, N_mass, C : N and Chl_mass (Figure 4) and mainly driven by different light regimes. This axis was significantly negatively correlated with TB of each species, but the relationship was stronger for R. pseudoacacia (Figure 5). The second axis, accounting for 24.5% of data variation, was mainly affected by A_max, Chl a : b and Chl : N, but not significantly linked to TB. Just as the three-way ANOVAs revealed, R. pseudoacacia and S. japonica responded differently to increased light. Under the low-light regime, the two species clustered in the ordination space, while under the high-light regime, they were clearly separated along the second axis (Figure 4).

Figure 4.

Principal component analysis of R. pseudoacacia and S. japonica across two light regimes based on 10 functional traits. See Table 1 title for trait abbreviations. The asterisks (*) denote the high-light regime.

Figure 5.

Linear regressions of TB against axis 1 score from PCA for R. pseudoacacia and S. japonica. LL, low light; HL, high light.

Discussion

Performance of R. pseudoacacia and S. japonica across different light regimes

Our study showed that in the low-light regime, R. pseudoacacia and S. japonica did not differ significantly in performance (indicated by TB); increased light stimulated the growth of both species, but R. pseudoacacia exhibited improved performance than S. japonica in the high-light regime. These results indicated that the exotic species, R. pseudoacacia, was better able to take advantage of increased light resource and enhance its performance. Paquette et al. (2012) reported a similar result with the invasive species Acer platanoides and a native congener. When resource availability increases, invasive species often exhibit a larger response, thus obtaining a higher performance than native species (Leishman et al. 2010), which is a prevalent ability among successful invaders (Richards et al. 2006). This is a possible mechanism of the fluctuating resource theory, which states that a plant community becomes more vulnerable to invasion with increasing resource availability (Davis et al. 2000, Dawson et al. 2012). Compared with co-occurring native species, there seem to be few invasive species that are always able to maintain fitness advantage across variable environments (Daehler 2003), therefore worldwide successful invasive species are inclined to be those better able to capitalize on increased resources, but not those ‘super invaders’, which are capable of maintaining greater performance across different environments (Dawson et al. 2012). In contrast, the native species S. japonica was less responsive to increased light intensity, which may be a fitness homoeostasis (Davidson et al. 2011).

Functional difference between R. pseudoacacia and S. japonica

In the high-light regime, R. pseudoacacia exhibited significantly higher performance and different functional trait values relative to S. japonica, which was consistent with the phenotypic divergence hypothesis: invasive species will be more successful in native communities if they show different functional traits from co-occurring natives. However, in the low-light regime, traits were similar between the two species (except for Chl_mass and Chl a : b), which might also indicate that the functional difference between the two species was habitat-dependent. Our results were consistent with the review of Daehler (2003), who synthesized numerous studies and concluded that the difference between invasive and native species largely relied on environmental conditions; under low resource availability, many functional traits of native species were equal to those of invaders or even superior. Based on the comparison between a large number of exotic and native species across a wide range of habitats, Tecco et al. (2010) also obtain similar results: invasive woody species exhibited significantly different functional attributes (more acquisitive) under favorable conditions, but in stressful habitats, they did not. Nevertheless, the phenotypic divergence of invaders and natives may be robust when individual comparisons are averaged across different environments (van Kleunen et al. 2010). Tecco et al. (2010) showed that while considering favorable and stressful conditions together, invasive woody species still displayed more acquisitive functional traits than natives. Similarly, in our study, when averaged over all treatments, R. pseudoacacia was still different from S. japonica in many functional traits, analogous to the case of performance, which indicated the habitat-dependence of functional difference and robustness of the phenotypic divergence between the two species.

In the high-light regime, R. pseudoacacia exhibited significantly higher TB and different functional trait values (A_max, RSR, N_mass and C : N) versus S. japonica, while in the low-light regime, TB and these functional traits were similar between the two species. This indicated that these functional traits might be closely related to the performance of R. pseudoacacia since they varied accordantly (Chen et al. 2013). Compared with natives, the higher growth rate of invasive species is considered as a key factor contributing to invasion success, which enables invaders to better acquire and utilize resources (James and Drenovsky 2007). In the high-light regime, R. pseudoacacia showed greater maximum net photosynthetic rate (A_max) and leaf N concentration (N_mass), a typical characteristic of invasive species in accordance with the theory of leaf economics spectrum that high A_max demands high N_mass and drives rapid growth (Wright et al. 2004). Robinia pseudoacacia also exhibited another prevalent trait of invaders, lower RSR (Godoy et al. 2012), through which R. pseudoacacia was able to achieve a larger canopy and capture light resources more efficiently.

Researchers have also tried to interpret the success of invasive species with the basic principle of ecological stoichiometry (González et al. 2010). The growth rate hypothesis deems that organisms can alter their C : N : P stoichiometry with the change of growth rate during development (Elser et al. 2003); plants characterized by high growth rates and great reproductive efforts generally have large nutrient requirements and thus low C : nutrient ratios (González et al. 2010). Therefore, in the high-light regime, the significantly lower C : N of R. pseudoacacia relative to S. japonica may indicate that R. pseudoacacia had a larger demand for nutrients and a higher growth rate, which contributed to a greater performance. Based on a large data set, Penuelas et al. (2010) also revealed that exotic species had significantly lower C : N than natives, regardless of whether the invasiveness of exotics was low or high. In addition, the leaf C : nutrient ratios are considered related to the capacity of C assimilation of plants per unit nutrient acquired, and thus to reflect nutrient-use efficiency to some extent. Hence, our data might also indicate a lower N-use efficiency of the exotic species R. pseudoacacia, consistent with the results of Scharfy et al. (2011). The possible explanation is that in favorable conditions, invasive species commonly adopt a more acquisitive and luxurious strategy to achieve a greater photosynthetic capacity and faster growth (Tecco et al. 2010).

Phenotypic plasticity in relation to light increase

Revealed by the reaction norm, across different light regimes, the plasticity of the performance trait was higher for R. pseudoacacia, which conferred it a significantly greater TB in the high-light regime. In the low-light regime, R. pseudoacacia accumulated similar biomass relative to S. japonica, therefore we can conclude that due to the superior ability to capitalize on increased light availability, R. pseudoacacia adopted a master-of-some strategy in response to the light gradient. Through this kind of strategy, invasive species can take better advantage of increased resources and increase fitness to a larger extent (Richards et al. 2006). Similarly, Mcalpine et al. (2008) reported that the survival of an invasive species conformed to the master-of-some scenario in response to a light gradient. Invaders in the study of Godoy et al. (2011) were also consistent with this scenario through plastic response of biomass to a nutrient gradient. For successful invaders, stress tolerance and resource capture are two crucial strategies, but the former is believed to have a tradeoff against growth and competitiveness, driven by resource limitation (Muth and Pigliucci 2007, Koehler et al. 2012, Chen et al. 2013). Only few vigorous species have both great stress tolerance and ability of resource acquisition, which enable them to outperform natives in both stressful and favorable environments (Liu et al. 2012, Mozdzer and Megonigal 2012). In addition, it is noteworthy that for a given environmental gradient, different fitness traits may resemble different scenarios, while a particular fitness trait may also conform to different scenarios under variant environmental gradients (Richards et al. 2006).

Across varied environments, the change in performance of invasive species may be a consequence of plastic responses of a series of functional traits (Dawson et al. 2012). Reflected by reaction norms, in relation to light increase, R. pseudoacacia exhibited higher plasticity for A_max and lower plasticity for N_mass, C : N and RSR relative to S. japonica, while other functional traits did not show any difference in plasticity between the two species. These results suggested that differences in plasticity between invasive and native species may vary among different traits (Chen et al. 2013); we cannot conclude that R. pseudoacacia was more plastic than S. japonica. With a higher plasticity of A_max, R. pseudoacacia achieved a greater photosynthetic capacity in the high-light regime and thus a faster growth; previous studies also reported the high photosynthetic capacity of R. pseudoacacia in natural conditions (Arena et al. 2008, Zheng et al. 2011); with a lower plasticity of RSR, N_mass and C : N, R. pseudoacacia was able to achieve lower RSR and C : N and higher N_mass in the high-light regime, and thereby to satisfy the high light and nutrient requirements due to rapid growth. Therefore, it may be inferred that with increased light, compared with S. japonica, R. pseudoacacia showed different levels of plasticity in functional traits in order to achieve a suite of favorable trait values in the high-light regime, which contributed to a greater performance. Hence, under changing environments, different functional trait values of invasive species relative to natives may be more important for invasion success than trait plasticity (Godoy et al. 2011), which has been confirmed by Godoy et al. (2012) using the method of model simulation.

Trait coordination across different light regimes

In this study, we estimated the magnitude of phenotypic integration by the number of significant correlations among functional traits. Inconsistent with some previous studies, such as Osunkoya et al. (2010), the results showed that the strength of trait correlation was similar between the two species, therefore our data did not support that R. pseudoacacia exhibited a higher level of phenotypic integration that S. japonica. Nevertheless, we found that A_max of R. pseudoacacia was significantly correlated with all the other functional traits, while for S. japonica it was only associated with Chl a : b. Correlations between traits may be due to causality, such as A_max and N_mass, or coordination, which occurs when specific trait values function better as an integral (Reich et al. 2003). This just conforms to the perspective of phenotypic integration: an integrated phenotype may better adapt to environmental changes through the coordinated variations of functional traits, and thus acquire more resources and deal with environmental stresses more efficiently (van Kleunen and Fischer 2005). Hence, we consider that the higher relevance between A_max and other functional traits for R. pseudoacacia may indicate that these traits responded to the light change coherently, which enabled this exotic species to better take advantage of increased light availability, reduce the expense of maladaptive plasticity, and thereby obtain a greater photosynthetic capacity and performance.

Moreover, when functional traits of the two species were sorted in the same multivariate space, the TB of R. pseudoacacia showed a greater correlation with the first ordination axis than that of S. japonica. Although this relationship may not necessarily indicate causality, at least it possibly suggested the higher coordination among the functional traits of R. pseudoacacia, which made them function better as a combination, respond more quickly to environmental variation and contribute more to performance (Godoy et al. 2012). Osunkoya et al. (2010) also obtained similar results while comparing several invasive and native vine species.

In fact, our understanding of phenotypic integration and its role in plant performance is still scarce (Godoy et al. 2012). Empirical studies have demonstrated that phenotypic integration of plants may restrain the expression of phenotypic plasticity, but the mechanisms are still unclear (Gianoli 2003, Gianoli and Palacio-López 2009). Theoretically, considered as two important components of plant functioning—flexibility and accordance, both phenotypic plasticity and integration are deemed to promote fitness (Gianoli and Palacio-López 2009, Godoy et al. 2012). Faced with environmental changes, plants may trade off between the two strategies; their relative importance for plant fitness might vary depending on the magnitude of environmental stress. Therefore, further studies are still needed on the role of phenotypic integration in invasiveness and its coordination with trait plasticity and values.

Responses of R. pseudoacacia and S. japonica to increased simulated N deposition

In this experiment, simulated N deposition showed little impact on either species. Before the experiment, the available N in each pot was 401.6 mg on average, while for the 3 and 9 g N m⁻² year⁻¹ deposition levels, the amounts of N added during the experiment was 72.7 and 218.2 mg. Moreover, as N₂-fixing species, both R. pseudoacacia and S. japonica are capable of reducing atmospheric N₂ to available N for their own use through symbiotic rhizobia. Especially, it is estimated that R. pseudoacacia can fix 23–300 kg N ha⁻¹ year⁻¹ (Danso et al. 1995, Noh et al. 2010, Cierjacks et al. 2013), which is comparable to or much higher than the simulated N deposition rate in this experiment. Therefore, the short-term N addition may not significantly affect the growth of the two species. Nevertheless, there are perhaps several reasons for the lack of responses of the two species to the simulated N deposition, which still remain to be explored.

In general, due to its widely acknowledged invasiveness, the strategies of R. pseudoacacia for superior growth over natives still remain a matter of concern. In this experiment, we only measured functional traits at the end of the growth period, but ignored the traits that should be recorded during the growth, such as relative growth rate, leaf emergent rate and leaf lifespan, which may also contribute to the superior performance. For example, Yamashita et al. (2000) found that when transferred from low to high-light condition, the success of an invasive tree (Bischofia javanica) seedling was due to not only the higher plasticity of photosynthetic capacity, but also its greater leaf production rate than native species. Yazaki et al. (2015) showed that faced with water deficiency, the invasive tree B. javanica shed leaves more conspicuously than natives to reduce water consumption, and thus to retain a relatively high growth performance. Therefore, besides the aspects we mentioned, there may be other reasons for the superior performance of R. pseudoacacia.

In addition, only small seedlings with limited genetic sources were studied and the treatments lasted for only 3 months in this experiment; moreover, the greenhouse study could not simulate the complicated environmental factors that may affect plant growth in natural conditions. Therefore, our results should not be extrapolated arbitrarily.

Conclusions

In general, our study showed that the exotic species R. pseudoacacia achieved significantly higher performance in the high-light regime, but performed similarly relative to the native species S. japonica in the low-light regime, indicating that it followed a master-of-some pattern in response to the light gradient. As the light intensity increased, R. pseudoacacia exhibited significantly different plasticity versus S. japonica in some crucial functional traits (A_max, N_mass, RSR and C : N), which enabled it to achieve a suite of divergent but advantageous trait values in the high-light regime, and thus a higher growth rate and greater performance. Moreover, across different light regimes, we found a higher level of coordination between A_max and other functional traits (also between TB and functional trait ordination axis) for R. pseudoacacia, which enabled it to capture and utilize increased light availability more efficiently. Therefore, we conclude that across different light regimes, three aspects of functional traits contributed to the superior performance of R. pseudoacacia: significantly different trait values and plasticity, and greater trait coordination. In the future, to reflect the differences in functional traits between the two species accurately, long-term experiments and field studies with plants from certain sources are necessary.

Supplementary data

Supplementary data for this article are available at Tree Physiology Online.

Conflict of interest

None declared.

Funding

National Science Foundation of China (Nos 31270374, 31470402, 31400173); Ministry of Science and Technology of China (No. 2011FY110300).