Competitiveness of the Exotic Silphium perfoliatum against the Native Urtica dioica : A Field Experiment

. Silphium perfoliatum (cup plant) is native to North America and is increasingly used as a bioenergy crop in Germany. Spontaneous occurrences of this species have already been detected in several European countries. To assess the possible risk to biodiversity by spreading of S. perfoliatum , we investigated the competitiveness of this species against the native and highly competitive Urtica dioica over four years in a feld experiment in Bayreuth (Germany). S. perfoliatum grew well among U. dioica , although its biomass was strongly reduced by surrounding U. dioica . Projection area, plant height, and reproductive potential were less or similarly reduced by surrounding U. dioica as by the intraspecifc competition. Moreover, S. perfoliatum signifcantly suppressed the growth of the competitive U. dioica . A settlement and establishment of S. perfoliatum in the native fora of Central Europe and a suppression of uncompetitive plant species are therefore conceivable.


Introduction
Invasive species are one of the major factors promoting global species extinction and the loss of biodiversity [1]. In the European Union, 39 invasive plant species are known and tried to be controlled [2]. It would probably be more successful to identify and manage risky species before they become invasive. Many of today's invasive plant species in Europe were introduced intentionally as ornamental plants or crops [3,4]. However, an investigation of the potential invasiveness of exotic plant species is hardly executed before they are cultivated on a large scale [5].
One such intentionally introduced species for which there are hardly any studies on the possible invasive potential is Silphium perfoliatum L. (cup plant). Native to eastern North America, it was introduced to Europe in the 18 th century as an ornamental plant [6]. It has been used as an alternative bioenergy crop in Germany since 2004 [7]. 2 EU-Regulation No. 1143/2014. To pursue the question of an invasive potential and especially the possible threat to biodiversity posed by S. perfoliatum in Central Europe, we investigated the competitiveness of this species for the frst time over four years in a feld experiment in the Ecological Botanical Gardens of the University of Bayreuth in Germany. Te central questions were as follows: (i) Initial phase (frst to second year of growth): (1) Is S. perfoliatum able to settle among native plants?
(ii) Establishment phase (second to fourth year of growth): (1) Is S. perfoliatum able to establish among native plants? (2) Is S. perfoliatum able to suppress native plants? (3) Is S. perfoliatum able to reproduce among native plants?
As the confronted native plant species, we chose Urtica dioica L. (common nettle). It is widely spread in Central Europe and belongs to the Urticaceae family [18][19][20]. Like S. perfoliatum, U. dioica is a perennial herbaceous plant that grows tall and is very competitive [19][20][21]. It occurs in nutrient-rich habitats and develops dominance stocks that are one of the most common fringe communities in Central Europe [20]. So far, S. perfoliatum settles predominantly in the immediate surroundings of its agricultural felds, which are mostly nutrient-rich fringes [13]. Terefore, S. perfoliatum can potentially coexist with U. dioica.
Assessing the competitiveness of the exotic S. perfoliatum against the competitive native U. dioica provides valuable insights that are relevant for evaluating a possible threat to biodiversity by S. perfoliatum and its invasive potential.

Experimental Setup.
Te experiment was carried out from May 2019 to August 2022 as a feld experiment at the Ecological Botanical Gardens of the University of Bayreuth (Germany). Mean air temperatures over the growing seasons from April to August each year were between 13.7 and 15.8°C (Table 1). Precipitation sum during this period varied between 195 and 342 mm among the years.
On 11 March 2019, triple the number of seeds we needed plants for the experiment of S. perfoliatum (Metzler & Brodmann Saaten GmbH, Ostrach, Germany) and U. dioica (Jelitto Perennial Seeds, Schwarmstedt, Germany) were sown in the greenhouse (Figure 1). Two and a half weeks after sowing, 1.5 times the required number of seedlings that we needed for the experiment were pricked out and continued to be cultivated in the greenhouse. On 16 May 2019, the saplings were planted in the feld. Tis date is set as the start of the experiment. Only plants that appeared to be vital were pricked out and planted. Plants that were infested with pests and those that grew particularly large or small were not selected for the cultivation and the experiment. After planting in the feld, the plants were watered only for the frst two weeks. Te experimental setup included three treatments for each of the two species (S. perfoliatum and U. dioica): Control treatment without competition, surrounded by eight plants of S. perfoliatum and surrounded by eight plants of U. dioica ( Figure 2). Each treatment was repeated 9 times (n � 9). Tey were planted in three blocks of two rows each ( Figure 3). In each block, the treatments in which the surrounding species were the same were placed next to each other. Two adjacent plots shared three surrounding plants ( Figure 3). Te order of the treatments and the order of the central species were randomly chosen, resulting in n � 3 per block. One individual of S. perfoliatum surrounded by U. dioica was much less vigorous from the frst year on and died in the fourth year. Tis individual was excluded from the analysis in all years, resulting in n � 8 for this treatment.

Data Collection.
At the end of each growing season, we surveyed growth parameters of the central plants (Table 2, Figure 1). As plant height, we measured the maximum height by calculating the mean of the fve highest shoots. Projection area was calculated as an ellipse with A � πab/4, where a is the measured maximum diameter of the projection area and b the perpendicular diameter to it [22]. Te number of living shoots taller than 15 cm was counted for each plant of S. perfoliatum. Aboveground biomass was harvested, dead biomass was removed, and the fresh weight of living biomass was measured with a scale (PM 4600 Delta Range, Mettler-Toledo GmbH, Greifensee, Switzerland, same scale for all further weight measurements). A representative subsample of at least one-third of the plant was taken and its fresh weight was measured. Tis subsample was dried in an oven at 90°C until the weight was constant. Dry weight was measured and extrapolated to the total living aboveground biomass per plant. Te number of capitula of S. perfoliatum was counted in the subsample from the stage of full fowering on (�fully expanded ray forets or later stages) and extrapolated to the total capitula per plant. In the second year, we harvested three ripe capitula per plant of S. perfoliatum, counted the number of fruits per capitulum, and calculated the mean of the three capitula.

Data Analysis.
Statistical analysis and data visualization were performed with R version 4.2.2 [23]. We used linear models (LMs) and checked the following assumptions using  [24,25]. In case of nonnormal distribution or heteroscedasticity of residuals, we transformed the parameters or used generalized linear models (GLMs). Log-transformation was executed with the natural logarithm. We checked the infuence of block ( Figure 3) with a LM, respectively, a GLM. Because it was not signifcant in each case, we eliminated the block for fnal models. We extracted the p values of the parameters in models with the "Anova" function with the F-test statistics of "car" package [26]. Signifcant diferences between the treatments were identifed with the Tukey's post hoc test on the models ("glht" function of "multcomp" package [27]). For diferences within one year, it was applied to separate univariate models. Te level of signifcance was always 0.05. As a measure of competition, we used the relative neighbor efect (RNE) [28], which is calculated as follows (1): In equation (1)   Year of growth

Growth of Silphium perfoliatum in the Initial Phase.
Te growth of S. perfoliatum in the initial phase (frst and second year of growth) was strongly afected by the treatment and the age of the plants (Table 3). In the control treatment (without competition), plants had the highest living aboveground biomass in both years: 409 ± 84 g (mean ± standard deviation) in the frst and 2141 ± 665 g in the second year of growth ( Figure 4). Biomass was signifcantly reduced by intraspecifc competition by 74% in the frst and 86% in the second year ( Figure 4, Table 4). However, competition by the surrounding U. dioica generated a signifcantly higher reduction by about 90%. Te strong increase in biomass from the frst to the second year of growth is due to the growth strategy of S. perfoliatum. In the frst year, it usually develops only a rosette of leaves and in the second year upright fowering shoots. In the present study, however, seven of the 26 plants developed one shoot already in the frst year (one in the control treatment and three in each competition treatment). In the second year, each plant developed shoots.

Growth of Silphium perfoliatum in the Establishment
Phase. Plant height of S. perfoliatum in the establishment phase (from the second year on) was the most afected by precipitation sum during the growing seasons from April to August (Table 5). Control plants were in the fourth and driest year 1.9 ± 0.2 m high (mean ± standard deviation) and in the third and wettest year 2.8 ± 0.2 m high ( Figure 5(a)). Treatment had also a signifcant efect on plant height (Table 5), although the competitive efects were low (Table 6). Intraspecifc competition led only in the fourth year to a signifcant reduction of plant height (21% compared to control plants). Surrounding U. dioica reduced plant height signifcantly in each year by 11 to 16%. Tis competitive efect was only in the second year signifcantly higher than by intraspecifc competition. Calculated over the whole establishment phase, the suppression of plant height by surrounding plants was not signifcant (Table 6). Additionally, plant height was slightly but signifcantly positively afected by the year of growth (Table 5). Nevertheless, there was no increasing plant height with increasing plant age measured due to the stronger efect by precipitation sum. Projection area of S. perfoliatum in the establishment phase was also the most afected by precipitation sum (Table 5). Projection area of the control plants ranged from 1.2 ± 0.5 m 2 in the fourth, driest year to 5.0 ± 2.5 m 2 in the third, wettest year ( Figure 5(c)). Treatment had also a signifcant efect on projection area (Table 5). Intraspecifc competition reduced projection area by 58 to 89% compared to the control plants (Table 6). Te reduction by interspecifc competition by U. dioica was similar and ranged from 52 to 83%. Troughout the establishment phase, suppression of projection area by competition was signifcant, but independent of the surrounding plant species (Table 6). However, there was a signifcant interaction between treatment and precipitation ( Table 5). Te suppressive efect of surrounding species on projection area increased with decreasing precipitation with both surrounding plant species, shown by higher RNE's in years with higher precipitation ( Table 6).
Number of shoots per plant of S. perfoliatum in the establishment phase was most afected by the treatment (Table 5). Control plants developed between 22 ± 4 (second year) and 38 ± 10 (fourth year) living shoots per plant ( Figure 6(a)). Surrounding by S. perfoliatum and U. dioica  led to a strong but not signifcantly diferent reduction of shoot number by 77% (by S. perfoliatum), resp. 86% (by U. dioica) in comparison to the control in mean over the years (Table 6). Te year of growth also had a signifcantly positive efect on shoot number. With increasing age, the number of shoots increased (Table 5, Figure 6(a)). Te efect of precipitation on shoot number was also signifcant, but much lower than of the year of growth (Table 5).

Growth of Urtica dioica in the Establishment Phase.
Plant height of U. dioica in the establishment phase was, as of S. perfoliatum, the most afected by precipitation sum during the growing seasons from April to August (Table 5). Control plants were the smallest in the fourth and driest year with 1.5 ± 0.1 m (mean ± standard deviation) and the tallest with 2.4 ± 0.3 m in the third and wettest year of growth ( Figure 5(b)). Treatment and its interactions with  (1). Te higher the RNE is, the higher is the competitive efect. Bold type indicates that the parameter in the given treatment is signifcantly lower than the control; * indicates that the parameter in the given treatment is signifcantly lower than in the treatment with the other surrounding species (Tukey's post hoc test on LM with log-transformed living aboveground dry biomass).  6 International Journal of Ecology precipitation and year of growth had also signifcant efects on plant height (Table 5). Intraspecifc competition did not signifcantly reduce plant height in any year (Table 6). In contrast, the plant height of U. dioica was signifcantly reduced by 26% by surrounding S. perfoliatum in the third and precipitation-richest year of growth. In the other and drier years, surrounding S. perfoliatum did not suppress the plant height of U. dioica.
Projection area of U. dioica was similarly afected by treatment as by precipitation (Table 5). Projection area of the control plants ranged from 1.2 ± 0.5 m 2 in the fourth to 4.5 ± 1.6 m 2 in the third year of growth and increased with increasing precipitation sum ( Figure 5(d)). Intraspecifc competition signifcantly reduced the projection area by 81 to 93% compared to control plants in each year. Suppression of projection area by surrounding S. perfoliatum was lower   (1). Te higher the RNE is, the higher is the competitive efect. Bold type indicates that the parameter in the given treatment is signifcantly lower than the control. but increased with time, which is refected in a signifcant interaction of treatment and year of growth. In the second year, there was no signifcant diference of projection area between the control plants and the plants surrounded by S. perfoliatum, whereas in the fourth year the reduction of the projection area of U. dioica by surrounding S. perfoliatum was similar to intraspecifc competition and amounted to 82% compared the control plants.

Reproductive Potential of Silphium perfoliatum.
Number of capitula per plant of S. perfoliatum was strongly afected by precipitation sum during the growing seasons from April to August and by the treatment (Table 5). In the fourth and driest year, control plants developed the fewest capitula with 82 ± 74 per plant (mean ± standard deviation) and in the third and wettest year the most with 718 ± 207 capitula per plant ( Figure 6(b)). Control plants had the most capitula in each year. Suppression by surrounding S. perfoliatum was very high and amounted to 92 to 100% reduction of the number of capitula compared to the control ( Table 6). Suppression by surrounding, U. dioica was signifcantly lower and between 82 and 89%. Te year of growth had also a signifcantly but lower positive efect on the number of capitula. Te number of fruits per capitulum was 30.8 ± 5.4 in the second year of growth and was independent of the treatment (LM, p � 0.362, Adjusted R 2 � 0.00, and n � 26).

Settlement of Silphium perfoliatum among Native Plants.
Te competitiveness of the exotic S. perfoliatum against the native U. dioica was investigated over four years in a feld experiment. We considered the frst two years as initial phase where settlement takes place.
In the initial phase, growth of S. perfoliatum, measured as living aboveground biomass was strongly reduced by competition; both with intraspecifc competition and with interspecifc competition by U. dioica. However, U. dioica had a stronger suppressive efect on S. perfoliatum (90-91% biomass reduction compared to the control plants) than S. perfoliatum to itself (intraspecifc competition with 74-86% reduction of biomass). Nevertheless, all suppressed plants survived the two years. All were vital and vigorous except for one individual. Tis one S. perfoliatum growing surrounded by U. dioica was less vigorous from the frst year on. In the fourth year, this individual died. We strongly suppose that the reason was root damage and not competition, because this plant was weak from the beginning on.
Usually, S. perfoliatum develops only a leave rosette in the frst year of growth and no shoots [6,29,30]. In the present study, a few individuals already developed one shoot in the frst year, mainly under competition. Tis was also observed in the feld experiment by Ende et al. [31]. We assume that in both studies, reasons were the early sowing and the precultivation under optimal conditions in the greenhouse before planting them out in the experimental sites in spring. Under more natural conditions, the plants would germinate and grow later in spring, so their growing season is shorter and the development of shoots in the frst year of growth is not to be expected.
In the present study, S. perfoliatum was thus well able to grow among the highly competitive U. dioica. However, settlement requires successful germination and seedling development. Tis was not investigated in the present study, because S. perfoliatum was planted as saplings among plants of the same age of U. dioica. It is known that seedlings of S. perfoliatum develop slowly and are therefore not very competitive in the frst weeks [32]. It was also observed that spontaneous colonization of S. perfoliatum took preferentially place in vegetation with about 25% open soil [13]. Additionally, this species requires full sun for optimal growth [6]. All these point to difculties for S. perfoliatum settling among dense and established native vegetation. However, there are apparently many suitable habitats for S. perfoliatum, especially in the areas around agricultural felds, because a study in northern Bavaria (Germany) recorded numerous spontaneous occurrences at such sites [13]. Moreover, the growth of S. perfoliatum has been demonstrated to be higher in moist soil conditions than in dry [31]. Tis makes successful settlement in moist habitats more likely [31].
We, therefore, assume that settlement of S. perfoliatum is possible among native plants, especially among weakly competitive plant species and in vegetation covers with disturbances and moist soil conditions.

Establishment of Silphium perfoliatum among Native
Plants. As establishment phase of S. perfoliatum, we defne the time from the second year of growth on, when this species usually starts to develop shoots and fowers [6,29,30].
Te plant height is an important parameter for competing plants, because it is decisive for the access of sunlight and thus for the rate of photosynthesis [33]. Te plant height of S. perfoliatum was hardly afected by surrounding plants from the second to the fourth year of growth in the present study. Much more afecting than competition for plant height was the precipitation sum during the growing seasons from April to August. Tis is in line with other studies where S. perfoliatum grew higher with higher precipitation or higher soil moisture [31,[34][35][36].
In contrast, the projection area of S. perfoliatum was strongly reduced by competition, similarly with S. perfoliatum and U. dioica as surrounding plants. Precipitation sum during the growing seasons had also a positive efect on the projection area of S. perfoliatum. Furthermore, the RNE-which is a measure for the reduction of projection area of S. perfoliatum by surrounding plants compared to the control plants-decreased with increasing precipitation. S. perfoliatum is therefore more competitive and more resilient with higher precipitation regarding the projection area.
Te number of shoots of S. perfoliatum was mainly reduced by competition by about 80% irrespective of the surrounding plant species. Interestingly, in contrast to the other two growth parameters, the number of shoots per plant was hardly afected by precipitation. More decisive was International Journal of Ecology the age of the plants. With increasing plant age, shoot number increased even under competition and even with decreasing precipitation. Tis is in line with the results of another experiment where the number of shoots of S. perfoliatum was also not afected by soil moisture [31]. Bury et al. [34] and Boe et al. [37] confrmed the correlation between shoot number and plant age. Te fact that S. perfoliatum develops less shoots under denser populations is confrmed in several studies [34,37,38].
Although, the growth of S. perfoliatum was reduced by competition, U. dioica did not suppress S. perfoliatum more than S. perfoliatum itself. According to Weber [20], U. dioica is able to develop dominance stocks, where other plant species have hardly a chance to grow. However, in the present study, S. perfoliatum grown among U. dioica was still vital and productive. S. perfoliatum is therefore apparently able to establish among native plants once it has settled there, especially with high precipitation and on moist soils. Among less competitive plant species than U. dioica, an establishment of S. perfoliatum is even more likely.

Suppression of Native Plants by Silphium perfoliatum.
With the present study, we not only intended to investigate the potential of S. perfoliatum to settle and establish among native vegetation but also aimed to address the important question whether S. perfoliatum could suppress native plants. For this purpose, we considered the years two to four of the feld experiment.
Te plant height of U. dioica was primarily afected by the precipitation sum during the growing seasons from April to August. Competition by surrounding plants hardly reduced the plant height of U. dioica, no matter whether it was intraspecifc competition or interspecifc competition by S. perfoliatum. Only in the one year with high precipitation, S. perfoliatum signifcantly reduced the plant height of U. dioica by 26% compared to the control plants. Tis shows once again that S. perfoliatum benefts from soil moisture and can thus exert competitive pressure on neighboring plants. In a pot experiment in Germany, the plant height of U. dioica was reduced signifcantly but similarly by intraspecifc competition as by competition by the exotic Impatiens glandulifera Royle [39]. Tese diferent results of an intraspecifc competitive efect are probably due to the diferent experimental setups (pot vs. feld) and the associated diferent conditions in terms of space and water.
Te projection area of U. dioica was reduced by about 80% in each year due to intraspecifc competition. When S. perfoliatum was the surrounding species of U. dioica, the competitive efect increased with plant age. In the second year, there was no signifcant efect, whereas in the fourth year, the surrounding S. perfoliatum reduced the projection area of U. dioica by 82% compared to the control plants. Tis was a similar efect as by intraspecifc competition in this year. It remains unclear whether this trend continues over further years. However, S. perfoliatum can persist for several decades [6], which could lead to a very high suppressive efect.
It is known that U. dioica is very competitive [20,21]. Because S. perfoliatum was able to restrict its growth, we assume that other less competitive native plant species would be more suppressed by S. perfoliatum. Tis efect could become stronger with increasing plant age of S. perfoliatum, as well as in years with high precipitation and in habitats with high soil moisture.

Reproductive Potential of Silphium perfoliatum among
Native Plants. Rhizome fragments of S. perfoliatum can serve as diaspores, and thus it enables a vegetative reproduction if the rhizomes become split [6,40]. In the present study, the potential of generative reproduction of S. perfoliatum under diferent competition treatments was investigated. It is known that the fowers of S. perfoliatum are visited by insects and fertile seeds are developed-also in Central Europe [9,10,30,41].
In the present study, S. perfoliatum developed about 30 fruits per capitulum. Tis was independent of the competition treatment. However, the number of capitula per plant of S. perfoliatum was signifcantly and strongly reduced by competition. With intraspecifc competition, the number of capitula of S. perfoliatum was reduced by 92 to 100% compared to the control plants. U. dioica as surrounding species caused a reduction of 82 to 89% that was signifcantly lower than by S. perfoliatum to itself. Besides, the number of capitula signifcantly increased with the precipitation sum during the growing seasons from April to August. Tis is in line with other studies where the number of capitula increased with increasing soil moisture [31,42]. Te number of capitula was also signifcantly positively afected by the age of the plant. With sufcient precipitation, the reproductive potential therefore could increase over the years.
Tus, the generative reproductive potential of S. perfoliatum is severely restricted under competition and more so under dry soil conditions than under moist ones. Nevertheless, even if S. perfoliatum produces only a few capitula, it can reproduce generatively due to the large number of fruits per capitulum.

Conclusions
Te present study is the frst to investigate the competitiveness of the introduced S. perfoliatum against native plant species in Central Europe. We used U. dioica as a native model species because it is vigorous, competitive, and prefers nutrient-rich habitats similar to S. perfoliatum [13,[19][20][21].
Te growth of S. perfoliatum was strongly reduced by competition, except the parameter plant height. However, it still grew and developed well especially in years with high precipitation. We assume that S. perfoliatum can settle and establish in the native fora of Central Europe. Furthermore, S. perfoliatum has a high competitive efect, especially with increasing plant age and with high precipitation because it was able to reduce the growth of the highly competitive U. dioica. A suppression of less competitive plant species is therefore conceivable. S. perfoliatum thus has both a high competitive efect and a high competitive response. It is known that low-competitive species are often valuable for nature conservation. A suppression of these species by S. perfoliatum could lead to a threat to biodiversity, so that S. perfoliatum could be classifed as "invasive" (Article 3, No. 2 EU-Regulation No. 1143/2014).

Data Availability
Te data are available in Table S1 in the Supplementary Materials.

Conflicts of Interest
Te authors declare that there are no conficts of interest regarding the publication of this study.