High-density native-range species affects the invasive plant Chromolaena odorata more strongly than species from its invasive range

Invasive plant species often form dense mono-dominant stands in areas they have invaded, while having only sparse distribution in their native ranges, and the reasons behind this phenomenon are a key point of research in invasive species biology. Differences in species composition between native and invasive ranges may contribute to the difference in distribution status. In this study, we found that the high-density condition had a more negative effect on C. odorata than the low-density condition when co-grown with neighbor plants from its native range in Mexico, while this pattern was not in evidence when it was grown with neighbors from its invasive range in China. Different competitive ability and coevolutionary history with C. odorata between native-range neighbors and invasive-range neighbors may lead to the inconsistent patterns.


Results
When C. odorata grown together with native species, density and species had significant effects on aboveground biomass, while category had no but category × density interaction had significant effect on the aboveground biomass (Table 1). Chromolaena odorata produced lower biomass when grown with native species from Mexico than when grown with species from its invasive range in China (Fig. 1), while native species from Mexico produced more biomass than native species from China when grown with C. odorata (Fig. 1). For each category, aboveground biomass at low density was significantly higher than at high density.
Category and density had no significant effects on change in biomass, while the effect of their interaction and the effect of species were significant (Table 1). When C. odorata was grown with native species from Mexico, the decrease in its biomass was greater in high-density communities than in low-density communities (Fig. 2). However, this was not the case when C. odorata was grown with native species from China (Fig. 2). For native species from Mexico and China, there were no significant differences for change in biomass between high and low density communities (Fig. 2). The decrease of biomass of C. odorata was higher when it grown with native species from Mexico than when it grown with native species from China (Fig. 2).

Discussion
The biomass of C. odorata grown with native species from Mexico was affected to a greater extent at high density than at low density (Fig. 2). This pattern was not in evidence when it was grown with species from its invasive range in China (Fig. 2). This is consistent with our original hypothesis.
Evolutionary responses to plant-plant interactions play a large role in shaping plant communities 14 . The different coevolutionary histories of C. odorata and native species from Mexico and those from China might explain these findings. In Mexico, Tithonia tubiformis, Chrysanthemum leucanthemum, and Eupatorium ligustrinum have longer coevolution history with C. odorata, and have adapted its' competitive strategies (growth rate, canopy shape, resource use, etc) accordingly. Further, as these three species are dominant species in Mexico, they may have evolved a stronger ability than C. odorata to cope with high-density conditions [15][16][17] . Therefore, when grown in high density with C. odorata, they may give a greater effect on C. odorata than in low density condition.
Meanwhile, C. odorata was introduced to China in 1930s, and the coevolution history between native species from China and C. odorata has occurred for a comparatively very short time. Though Abutilon indicum, Bidens biternate, and Artemisia myriantha are dominant species, C. odorata might not be sensitive to their competitive strategies. Furthermore, some invasive species have "novel weapons", and native species from the invasive range are usually more vulnerable to these "weapons" owing to short coevolution history 10,18,19 . C. odorata produces odoratin (Eupatorium) (C 19 H 20 O 6 ), a potential novel allelochemical 3 . In a previous study, we found that seedlings Variable Category (C) Density (D) C*D Species (Category)  from China are more vulnerable to this allelochemical than seedlings from Mexico 3 . This might contribute to the successful invasion of C. odorata.
Previous studies also revealed invader-neighbor interactions differ between the native range and invasive range. In the native range, the impact of invasive Centaurea stoebe is mainly driven by competition for the same limiting resources between the invader and neighbors, whereas in the invasive range, other factors play an important role 11,20 . These factors include soil biota, exploitation of resources that are not used by neighbors, and special traits of the invasive species that directly interfere with competitors 11,20 . Different invader-neighbor interactions could also explain the inconsistent patterns in this study. Native species from Mexico are more competitive than native species from China (Fig. 1). When they grow together with C. odorata, the competition will begin earlier at high density than at low density. This will increase the advantage of native species and the disadvantage of C. odorata, which leads to greater decrease of biomass for C. odorata at high density than at low density. However, when C. odorata competes with native species from China, in addition to direct growth competition, novel weapon 3 , soil biota 21 , and herbivores 22 all influence the competition, which leads to the mixed results (Fig. 2).
In this study, all native species are common, dominant and have sympatric distirbution with C. odorata in each site, which made our results more plausible in understanding the mechanisms of invasion success of C. odorata. However, the results might also due to sampling effect (such as rapid growth rate, early establishment, etc) for only three species were used within each range. More species should be examined in future studies.
In conclusion, native-range species at high density negatively affected C. odorata to a greater extent than those at low density. This pattern was not observed when C. odorata was grown with species from its invasive range in China. Different species composition and geographical differences in coevolutionary histories between native-range plants and invasive-range plants might explain these results.

Methods
Seed collection. In 2012, we collected seeds of C. odorata and three native species from China (Abutilon indicum, Bidens biternata, and Artemisia myriantha) in Xishuangbanna and Puer ( Table 2). Seeds of three native species from Mexico (Tithonia tubiformis, Chrysanthemum leucanthemum, and Eupatorium ligustrinum) were collected in 2011 in Morelos and Veracruz ( Table 2). All these native species are common and dominant species in each site, and they can form dense stands in some patches. For each species, seeds were collected from more than 10 individuals growing at least 20 m apart, then mixed uniformly in paper bags. Seeds were germinated in a seedbed in April 2013 and transplanted to a common garden in June. In late 2013/early 2014, newly produced seeds of each species were collected from these next-generation plants in order to exclude maternal effect.

Common garden experiments. Common gardens were constructed in Xishuangbanna Tropical Botanical
Garden (21°6′N, 101°15′E, 560 m above sea level), Yunnan Province, southwest China. Here the mean annual temperature is 21.7 °C, the mean temperature in July (the hottest month) is 25.3 °C, and that in January (the coolest month) is 15.6 °C. The average annual precipitation is 1557 mm with a dry period lasting from November to April. The soil is latosol, and total nitrogen, phosphorus and potassium contents were 1.07 g/kg, 0.44 g/kg and 13.48 g/kg, respectively. Twenty-six 2 × 2 m artificial communities were constructed in early 2015 in each common garden, with each community 1 m apart from its neighbors. Three replicate common gardens were constructed. In March 2015, seeds of each species were sown separately into greenhouse seed beds. In the following June, similarly-sized seedlings were transplanted into artificial communities, consisting of monoculture of each species, either C. odorata + native range species or C. odorata + invasive range species, at either of two growing densities: high (36 individuals per plot) or low (8 individuals per plot) (Fig. 3). These densities were set according to our field observation. One of each type of artificial community was constructed in each of the three common gardens. At the first month, a few seedlings died and we substituted them with seedlings of similar size immediately.
In December 2015, the aboveground parts of all plants were harvested, oven-dried at 80 °C for 72 h, and weighed. The magnitude of the neighbor effect was evaluated in terms of the change in biomass, which was calculated as follows:  Table 2. Background information on seeds and collecting sites. where Biomass together represents the mean value of individual aboveground biomass for each species when grown together with other species, and Biomass monoculture represents the mean value of individual aboveground biomass for this species when grown in monoculture.

Statistical analysis.
The following four categories were defined: C. odorata co-grown with species from Mexico, native species from Mexico, C. odorata co-grown with species from China, and native species from China. When C. odorata was grown together with other species, differences in aboveground biomass and change in biomass among categories were determined by a linear mixed-effects model. Category, density and their interaction were treated as fixed factors. Species nested within category was treated as a random factor. A tukey HSD test was used to compare the differences of neighbor effects between high-and low-density communities. Duncan test was used to compare the differences of biomass among categories. We make a log transformation for aboveground biomass to suitable for the equality of error variances. All analyses were conducted using SPSS 18.0.
Data availability. The data generated during this study are available from the corresponding atuthor on reasonable request.