Responses of a native plant species from invaded and uninvaded areas to allelopathic effects of an invader

Abstract Invaders exert new selection pressures on the resident species, for example, through competition for resources or by using novel weapons. It has been shown that novel weapons aid invasion but it is unclear whether native species co‐occurring with invaders have adapted to tolerate these novel weapons. Those resident species which are able to adapt to new selective agents can co‐occur with an invader while others face a risk of local extinction. We ran a factorial common garden experiment to study whether a native plant species, Anthriscus sylvestris, has been able to evolve a greater tolerance to the allelochemicals exerted by the invader, Lupinus polyphyllus. Lupinus polyphyllus produces allelochemicals which potentially act as a novel, strong selective agent on A. sylvestris. We grew A. sylvestris seedlings collected from uninvaded (naïve) and invaded (experienced) sites growing alone and in competition with L. polyphyllus in pots filled with soil with and without activated carbon. Because activated carbon absorbs allelochemicals, its addition should improve especially naïve A. sylvestris performance in the presence of the invader. To distinguish the allelochemicals absorption and fertilizing effects of activated carbon, we grew plants also in a mixture of soil and fertilizer. A common garden experiment indicated that the performances of naïve and experienced A. sylvestris seedlings did not differ when grown with L. polyphyllus. The addition of activated carbon, which reduces interference by allelochemicals, did not induce differences in their performances although it had a positive effect on the aboveground biomass of A. sylvestris. Together, these results suggest that naïve and experienced A. sylvestris plants tolerated equally the invader L. polyphyllus and thus the tolerance has not occurred over the course of invasion.


| 6117
LYYTINEN aNd LINdSTRÖM and Oduor (2013)), but the absence of evolutionary responses has also been reported (Goergen, Leger, & Espeland, 2011;Lau, 2006;Mealor & Hild, 2007). Native populations can respond evolutionarily if they, for example, possess enough genetic variation, selection pressure is strong and consistent, and gene flow between uninvaded and invaded native populations is limited Strauss et al., 2006). In addition to the evolutionary factors, ecological factors, such as multiple species interactions (Lau, 2006), may limit the evolutionary capacity (review by Lau and terHorst (2015)).
Allelopathy refers to the negative and positive effects of biochemicals, so-called allelochemicals, produced by an organism on another organism (Rice, 1984). In the context of invasion, the role of allelochemicals has been studied mainly from the perspective of invaders as a potential explanation for their superiority in an invaded area. The novel weapon hypothesis states that resident plants are negatively affected by the biochemicals released by an invader (i.e., allelopathic effects) to which they are not adapted (Callaway & Aschehoug, 2000). The fate of resident plants is not necessarily a stable state if they can evolve to tolerate invaders' allelochemicals (Callaway, Ridenour, Laboski, Weir, & Vivanco, 2005), making coexistence with an invader more likely.
Alkaloids can also be released from dead plant tissue. In L. polyphyllus, the allelopathic effects of litter seem to be minor compared to that of the root exudates. Litter of L. polyphyllus can reduce germination and increase germination time of native species but on the other hand, it increases seedling biomass, potentially due to the nutrients released through decomposition (Loydi, Donath, Eckstein, & Otte, 2015).
Although L. polyphyllus often forms monospecific stands, there are some species which are able to coexist with it such as the cow parsley, Anthriscus sylvestris (Valtonen et al., 2006). Our previous research showed that at least some A. sylvestris populations have evolved an increased competitive ability against L. polyphyllus (A. Lyytinen & L. Lindström, unpublished data). As native species are naïve to the allelochemicals produced by L. polyphyllus, the tolerance of remnant species to these allelochemicals may have increased in response to selection caused by L. polyphyllus. We set up an experiment to test whether A. sylvestris has adapted to allelochemicals released by L. polyphyllus into the soil. We collected A. sylvestris seedlings from uninvaded and invaded sites and grew them either alone or in competition with L. polyphyllus. We also manipulated the soil conditions by adding activated carbon, which absorbs allelochemicals. If local adaptation to invader chemicals has taken place, L. polyphyllus is expected to decrease the performance of A. sylvestris with no experience with L. polyphyllus compared to those conspecific individuals that have experienced invasion as they are naïve to the allelopathic chemicals of the invader. In the presence of activated carbon, this difference in performances is expected to be reduced because activated carbon should increase especially the performance of the most susceptible individuals. To assess the allelochemicals absorption and fertilizing effects of activated carbon, plants were also grown in a soil mixed with fertilizer.  (Table 1). Activated carbon (20 ml/L of soil, Merck KGaA) was mixed with the substrate to ameliorate the allelopathic effects of L. polyphyllus. As a control for the activated carbon treatment, we added fertilizer (0.4 g per pot, Kekkilä, puutarhalannoite, NPK 9-4-13) to another set of pots. This allowed us to distinguish the absorption of allelochemicals and fertilizing effects of activated carbon F I G U R E 1 A mixed population of Anthriscus sylvestris and the invader Lupinus polyphyllus. (Weißhuhn & Prati, 2009). Pots were placed randomly in a common garden at the University of Jyväskylä. The height of the stem was measured every second week until the growth was levelled off. After each measurement, the order of pots was again randomized. After 14 weeks, the stem was cut at the height of 3 cm from ground level, dried at 48°C for 7 days and weighed.

| MATERIAL S AND ME THODS
For overwintering, pots were dug into the soil so that the pot rim was at the soil level. In the following spring, pots were dug up and placed in a common garden. Only those pairs where both plants were alive were included in the experiment (Table 1). Similarly, as in the first growing season, the height of the stem was measured every second week until the growth was levelled off when the stems and roots were harvested. After drying at 48°C, the stems and roots were weighed. In total, the experiment lasted 448 days (two growing seasons).

| Data analysis
We performed separate tests for competition treatments (grown alone, grown together with L. polyphyllus). To test whether plant height growth pattern of A. sylvestris differs with the invasion history or soil treatment in the first growing season, we performed a repeated 2-way ANCOVA with the height measurements as different factor levels and the invasion history (naïve, experienced) and soil treatment (control, activated carbon, fertilizer) as fixed factors and a root length of a seedling as a covariate. For the data from the second growing season, the model included also the initial shoot height of the seeding as a covariate. To test which of the factors affect the final shoot and root dry weight, and root-shoot ratio, we performed separate Generalized Linear Models where the invasion history and soil treatment were fixed factors, and a root and shoot length of a seedling were covariates. The pairwise comparisons were tested with least significant difference (LSD). Survival from the beginning of the experiment to the following spring was analyzed with Binary logistic regression. All analysis were performed with SPSS Statistics 24.  Table 2). The addition of fertilizer resulted in a higher grown pattern compared to the control treatment (LSD: p = 0.017 and p = 0.006, respectively). The addition of activated carbon did not induce differences compared to fertilizer (p = 0.072 and p = 0.199, respectively) or control treatments (p = 0.505 and p = 0.142, respectively). The effect of the soil treatment did not differ with the invasion history. The root or shoot length of a seedling did not affect the growth pattern.

| The final biomass and root-shoot ratio in A. sylvestris
In the first growing season, the invasion history by soil treatment interaction and the main effect of invasion history (naïve, experienced) on shoot biomass in alone grown A. sylvestris were not significant (Table 3, Figure 3a). In contrast, shoot biomass differed with the soil treatment. Activated carbon increased the shoot biomass by 42%  (Table 3). Only the shoot length of the seedling affected shoot biomass and the root-shoot ratio. The longer the shoot was, the heavier the shoot biomass and the smaller the ratio were.
When A. sylvestris plants were grown in competition with L. polyphyllus for one growing season, the invasion history by soil treatment interaction and the main effect of invasion history on shoot TA B L E 1 The number of plants in each treatment at the beginning of the first and second growing season. In the second year, only those plant pairs were included where both were alive after the first winter. Plants were collected from uninvaded (naïve) and invaded (experienced) sites. Anthriscus sylvestris plants, which were grown alone or in competition with the invader L. polyphyllus, were allocated to soil treatments: control, activated carbon, and fertilizer  Figure 3b). On average, activated carbon increased the shoot biomass by 53% compared to the controls (p = 0.013), producing plants with equal biomass as the addition of fertilizer (p = 0.143). Fertilizer, however, increased biomass even more than the addition of activated carbon, by 86% compared to the controls (p < 0.001). Unlike the root length of the seedling, the size of the seedling affected shoot biomass. The longer the shoot length of the seedling was, the heavier the shoots were. The results from the second growing season were consistent with the first growing season. Neither the invasion history by soil treatment interaction nor the invasion history significantly affected the shoot and root biomass whereas soil treatment did. A. sylvestris plants, which grew in fertilized substrate, had a higher shoot (LSD: p < 0.001, increase 112%, Figure 3b) and root biomass (p < 0.001, increase 126%, Figure 3c) than control plants. The addition of fertilizer also produced A. sylvestris plants with greater root biomass than the addition of activated carbon (p = 0.001), but such difference was not observed in shoot biomass (p = 0.106). The shoot (p = 0.120) and root biomass (p = 0.372) of plants grown with and without activated carbon were equal. Root-shoot ratio was not affected by invasion history, soil treatment, their interaction, or covariates (the shoot and root length of the seedling, Figure 3d).  There are several potential explanations for the absence of differences in performances between naïve and experienced A. sylvestris.

| D ISCUSS I ON
First, the strength of the selection might not been strong enough to result in local adaptation (Kawecki & Ebert, 2004). In our experiment,  TA B L E 2 Results of the repeated 2-way ANCOVA testing for effects of the invasion history (naïve, experienced) and soil treatment (control, activated carbon, fertilizer) on the growth pattern in A. sylvestris plants which were grown alone or in competition with the invader L. polyphyllus. A shoot and root length of a seedling were covariates TA B L E 3 Results of General Linear Model testing for effects of the invasion history (naïve, experienced) and soil treatment (control, activated carbon, fertilizer) on the final biomass of the shoot and root, and root-shoot ratio in A. sylvestris grown alone or in competition with the invader L. polyphyllus. A shoot and root length of a seedling were covariates. A significant factor is marked in bold font its homogenizing effect has been stronger than the strength of selection (Kawecki & Ebert, 2004). As A. sylvestris is a common and abundant plant in Finland in overlapping areas with L. polyphyllus (Lampinen & Lahti, 2016) and its pollinators travel long distances (Rader, Edwards, Westcott, Cunningham, & Howlett, 2011), gene flow between naïve and experienced populations is likely. Third, it is also possible that invasion has been recent and there has not been enough time to allow evolutionary changes. Because of imperfect knowledge of invasion history of L. polyphyllus, we are able only to estimate the length of the period of association. The first reports of wild stands of L. polyphyllus in Finland are from 1895 (Fremstad, 2010) and they reached the present collection sites, Central Finland, by 1970s (Lampinen & Lahti, 2016).
Based on this, we can estimate the maximum age of L. polyphyllus populations to be 40 years. Other plants have been reported to adapt to novel allelochemicals even in a shorter time period (20-30 years) (Callaway et al., 2005) although the length of the association also has based on estimated invasion times.
The lack of differences among naïve and experienced A. sylvestris plants as a response to L. polyphyllus could also indicate that roots may not exude allelochemicals in large quantities to have measurable allelopathic effects (but see Wink, 1983). The majority of the studies examining alkaloids in Lupinus species has focused on the seeds. As alkaloids are synthetized mainly in the green parts of L. polyphyllus (Wink, Hartmann, & Witte, 1980), it is possible that adding withering leaves could have had larger effect on the growth of A. sylvestris.
Adding activated carbon to the substrate resulted in the improved performance of A. sylvestris plants when grown in competition with L. polyphyllus but only in the first growing season.
Furthermore, activated carbon increased the above ground biomass relatively more in plants grown in competition with L. polyphyllus (a 53% increase compared to controls) than that of those grown alone  (Ridenour & Callaway, 2001).
In the second growing season, the addition of activated carbon to the substrate did not alter the effects of the presence of L. polyphyllus on A. sylvestris, suggesting the absence of allelopathic effects.
One possible explanation for the dissimilar results compared to the first growing season is the reduced sample size. The mortality among the pairs which were exposed to competition from L. polyphyllus was high, diminishing the power of statistical tests to detect differences.
As A. sylvestris appeared to be relatively tolerant to L. polyphyllus allolechemicals, one might ask whether it has experienced competition from a species which also contains the same alkaloids as L. polyphyllus. The main quinolizidine alkaloid in L. polyphyllus is lupanine  which is not known to occur in any other species growing wild in Finland than L. polyphyllus (Aniszewski, 2007). Thus, naïve A. sylvestris has most likely not been exposed earlier to it.
Although we found some indications of allelopathic effects, we did not find evidences that experienced A. sylvestris plants have evolved an increased tolerance to the L. polyphyllus allelochemicals.
However, there was among individual variation in aboveground biomass, indicating potential for evolution. On the basis of a greater positive effect of addition of fertilizer compared to activated carbon, we can infer that competition for resources is more important factor than allelopathy behind the harmful effects of L. polyphyllus on A. sylvestris.

ACK N OWLED G M ENTS
We thank Agnese Comellato, Aigi Margus, and Jemina Nevala for assistance. This research project was supported by the Finnish Cultural Foundation, the Academy of Finland (project number 252411, Finnish Centre of Excellence in Biological Interactions Research) and Konnevesi writing grant from the University of Jyväskylä.

CO N FLI C T O F I NTE R E S T
Authors declare no competing interest.

AUTH O R CO NTR I B UTI O N S
AL and LL conceived the ideas and designed methodology; AL collected and analyzed the data; AL and LL led the writing of the manuscript. Both authors contributed critically to the drafts and gave final approval for publication.

DATA ACCE SS I B I LIT Y
When the paper will be published, our data will be included in the