Prey naïveté rather than enemy release dominates the relation of an invasive spider toward a native predator

Abstract Ecosystems may suffer from the impact of invasive species. Thus, understanding the mechanisms contributing to successful invasions is fundamental for limiting the effects of invasive species. Most intuitive, the enemy release hypothesis predicts that invasive species might be more successful in the exotic range than resident sympatric species owing to the absence of coevolution with native enemies. Here, we test the enemy release hypothesis for the invasion of Europe by the North American spider Mermessus trilobatus. We compare the susceptibility of invasive Mermessus trilobatus and a native species with similar life history to a shared predator with which both species commonly co‐occur in Europe. Contrary to our expectations, invasive Mermessus trilobatus were consumed three times more frequently by native predators than their native counterparts. Our study shows that invasive Mermessus trilobatus is more sensitive to a dominant native predator than local sympatric species. This suggests that the relation between the invasive spider and its native predator is dominated by prey naïveté rather than enemy release. Further studies investigating evolutionary and ecological processes behind the invasion success of Mermessus trilobatus, including testing natural parasites and rapid reproduction, are needed to explain its invasion success in Europe.


| INTRODUC TI ON
Nonindigenous species can play influential roles in their exotic range once becoming invasive. Invasions are considered successful when alien species establish and rapidly expand their ranges in novel environments by overcoming biogeographical barriers and ecological pressures (Sakai et al., 2001). The impact of invasive species on native ecosystems has been described since the middle of the 20th century (Elton, 1958). However, the mechanisms behind the often striking success of invasive species are still uncertain (Schultheis et al., 2015). Up to 39 hypotheses were developed to better describe the processes behind successful invasions (Enders et al., 2019). As one of the most predominant and intuitive, the enemy release hypothesis posits that nonindigenous species are released from the pressure of predators and parasites once introduced to their exotic ranges (Elton, 1958;Keane & Crawley, 2002). Introduced species | 11201 NARIMANOV et Al.
might have advantages over resident species against native enemies by, for example, not being recognized as prey or hosts for parasites in the exotic range (e.g., Cottrell & Shapiro-Ilan, 2003;Gozzi et al., 2020;Montes et al., 2020;Tierney et al., 2020). However, Elton (1958) mentions that species leaving coevolved predators and parasites from the native areas immediately meet novel potential enemies once introduced in the exotic range. Hence, due to the lack of the coevolutionary history with novel predators, parasitoids, and pathogens, introduced species might be naïve toward novel enemies' archetypes under comparable or even higher enemy pressure in their exotic range (Cox & Lima, 2006).
Despite the popularity of the enemy release hypothesis (Hierro et al., 2005), the growing literature provides only mixed support (Heger & Jeschke, 2014;Schultheis et al., 2015). Such a discrepancy might come from the studies' different approaches based on the scale of the analysis. Biogeographical studies investigate enemy release comparing invasive animals from native and exotic populations.
In contrast, community studies examine native and invasive species from the same community in the exotic range (Colautti et al., 2004).
While studies at the biogeographical scale largely support the enemy release hypothesis, the results from community studies are equivocal (Colautti et al., 2004). Such differences between biogeographical and community scale studies might arise due to, for example, failure to distinguish two types of enemy release, namely compensatory, when the limited resources utilized for defense are repositioned elsewhere, and regulatory, when the loss of enemies leads directly to increase in demographic parameters (Colautti et al., 2004). This may lead to inaccurate conclusions about the net effect of enemy release at biogeographical scales. Further, studies investigating enemy release comparing the number of enemy species between natural and exotic ranges at the biogeographical scale might be ambiguous since invasive species and their enemies are often better studied in their native rather than exotic range. Hence, more enemies would be expected in native populations due to sampling efforts (Colautti et al., 2004). Additionally, only a portion of the population is being relocated to exotic regions during the transport of invasive species. Therefore, introduced populations are often genetically less diverse compared with native populations. Hence, such invasion bottlenecks could also lead to nonvalid comparisons of the populations on the biogeographical scale (for a more extensive review, see Colautti et al., 2004). Consequently, all introduced species lose enemies at the biogeographical level, irrespective of their release from enemies in their introduced range at the community level (Colautti et al., 2004).
An increasing number of studies indicate a changing role of enemy release through the different invasion phases, namely introduction, establishment, and spread (first: Drake, 2003;reviewed in Roy et al., 2011). Accordingly, release from enemies might play different roles during the introduction, establishment, or spread of invasive species in their exotic range. For instance, the parasitism of invasive mosquito Aedes albopictus (Diptera: Culicidae) by a native enemy is low in the introduced area only for at least two years following the colonization (Aliabadi & Juliano, 2002). Still, many invasive species fail to establish in the exotic regions after introduction.
One of the most plausible contributing mechanisms of establishment failure of invasive species may be an increased pressure by novel enemies in the introduced range (Cox & Lima, 2006;Elton, 1958). Therefore, the enemy release hypothesis as a driving force behind successful invasions should be tested for already established invasive species that are in their spreading phase (i.e., abundant or dominant) in the exotic range. Furthermore, studies investigating the role of enemy release as a causal mechanism of invasiveness are mainly based on invasive plant and vertebrate species (e.g., Carpenter & Cappuccino, 2005;Gozzi et al., 2020;Hawkes, 2007;Hierro et al., 2005;Lankau et al., 2004;Liu & Stiling, 2006;Meijer et al., 2016;Montes et al., 2020;Schultheis et al., 2015;Tierney et al., 2020), whereas only a limited number is focused on arthropods (e.g., Aliabadi & Juliano, 2002;Juliano et al., 2010;Paula et al., 2021;Zuharah & Lester, 2010).
The North American dwarf spider Mermessus trilobatus (Araneae: Linyphiidae; formerly known as Eperigone trilobata; Millidge, 1987) was first recorded in Europe in the late 1970s near Karlsruhe in South-West Germany (Dumpert & Platen, 1985). The species has undergone a concentric range expansion in Europe, spreading by > 1,000 km in less than 50 years (Hirna, 2017) and often reaching high local abundances (Narimanov et al., 2021;Schmidt et al., 2008). The invasion success of M. trilobatus seems to be neither based on high competitiveness toward native linyphiids (Eichenberger et al., 2009) nor a ruderal strategy, as they do not benefit from soil disturbance (Narimanov et al., 2021). Therefore, reduced susceptibility to native predators might explain the invasion success of M. trilobatus in Europe.
Here, we investigate, at the community level, whether the invasion success of M. trilobatus in Europe is explained by the release from the pressure of native predators. Thus, we compare the invasive M. trilobatus and a native sympatric species' susceptibility to a shared native predator with which they frequently co-occur. We expect that in contrast to the shared coevolutionary history of the native prey and predator, invasive M. trilobatus would benefit from reduced predation by native predators, which could explain their invasion success in Europe.

| Study species
We chose Erigone dentipalpis (Araneae: Linyphiidae) as native prey because of their similar size (Table 1) and hunting mode to M. trilobatus and because the two species often dominate in the same habitats (Narimanov et al., 2021). Spiders are exposed to various natural enemies, including other spiders as perhaps the most important predators (Foelix, 2011). Therefore, we chose Pachygnatha degeeri (Araneae: Tetragnathidae; body length = 3-4.2 mm), the most abundant linyphiid-eating spiders, as predators for the experiments.
Pachygnatha degeeri are free hunters living close to the ground of the grasslands where both M. trilobatus and E. dentipalpis are found and can easily climb and invade linyphiid webs. Moreover, these generalist predators are not found in North America and, thus, are ideal candidates as native European predators (Nentwig et al., 2021). We sampled all spiders from perennial hay meadows as the preferred habitat of M. trilobatus (Narimanov et al., 2021). The meadows were situated next to the river Queich, close to Landau in Germany (see Table 2 for coordinates). Spiders were sampled between July and September 2020 using a vacuum sampler (modified STIHL SH86 blower; Stihl, Waiblingen, Germany). We sampled 14 M. trilobatus and 16 E. dentipalpis females and 85 adult P. degeeri individuals. All

| Experimental design
Experiments were performed in 405-ml glass jars with approximately 1 cm layer of moistened plaster on the bottom and five vertical sticks to facilitate web building. We used only adult linyphiids reared in the laboratory. Prior to experiments, we measured all spiders' prosoma widths (see Table 1 for means) as an estimate of body size that is independent of the current feeding condition (Moya-Laraño et al., 2008).
We assigned a randomly chosen pair of prey (invasive M. trilobatus and native E. dentipalpis) to the same predator (P. degeeri). Then, each linyphiid pair was tested with the same predator during two trials in random order. We calculated the difference in prosoma width by subtracting the respective value of the predator minus the prey. We let linyphiids build webs in the glasses without food for two days before experiments. All linyphiids, irrespective of sex and species, built a web. Simultaneously, we starved predators also for two days prior to each trial for standardization. From previous observations, we expected that E. dentipalpis builds the web slightly closer to the surface than M. trilobatus. As spiders in low webs may be more exposed to ground-hunting predators such as P. degeeri, we sprayed the webs with water and measured their lowest and the highest position to the plaster in each glass after two days of web building. We placed predators on the surface of the plaster, avoiding any damage to webs, and gave the trials three days. We checked spiders every 24 hr. In total, we had 202 trials and tested 101 M. trilobatus and 101 E. dentipalpis.

| Statistical analysis
We modeled the consumption rate (consumed, not consumed) by fitting a generalized linear mixed-effect model (GLMM) for a binomial response from the lme4 package (Bates et al., 2015) in R 4.0.3 (R Core Team, 2020). We then applied ANOVA chi-square test (the car package in R; Fox & Weisberg, 2019) to the GLMM model to analyze the effects of prey species (M. trilobatus, E. dentipalpis), prey and predator sex, the difference in prosoma width of the predator and prey, and web minimum and maximum positions to the surface on the consumption rate. We included predator ID as a random factor since each predator was used at least twice during experiments.
Additionally, we modeled the linyphiids' web positions to the surface (minimum and maximum) by fitting linear models (lm) from the R package stats (R Core Team, 2020) and included linyphiid species as fixed predictors. We then applied ANOVA F test to the lm models to investigate the web-building strategies of two species (M. trilobatus and E. dentipalpis). We validated the lm model results using permutation tests (PermTest function from pgirmess package in R; Giraudoux, 2021).

| RE SULTS
Opposite to our expectation, the invasive M. trilobatus was consumed almost three times more often compared with native E. dentipalpis (Table 3, Figure 1) with predators) were consumed with slightly a higher rate than larger ones (Table 3, Figure 2). There were no effects of spiders' sex (predators and prey) and linyphiids' web positions (minimum and maximum) on their susceptibility to predation (Table 3). However, on average, native E. dentipalpis built their webs around 2.5 times closer to the surface (plaster) than invasive M. trilobatus (web minimum; F 1, 200 = 9.843, p = .002; Figure 3). There was no difference in linyphiids' web maximum positions to the surface (web maximum; F 1, 200 = 2.472, p = .118). In total, 95 out of 202 linyphiids were consumed. The highest number of M. trilobatus was consumed during the first two days (35 and 24, respectively), followed by the last day (11).
Similarly, the highest number of E. dentipalpis (15) was consumed the first day, leaving the following two days with an equal number of individuals consumed (5 each). A higher number of females of E. dentipalpis compared with males were consumed (20 and 5, respectively), whereas similar numbers of M. trilobatus females and males were consumed during experiments (39 and 31, respectively). Our results show that the invasive M. trilobatus is sensitive toward local European predators. Mermessus trilobatus has undergone rapid concentric range expansion in Europe. Spiders used in our experiments were derived from the populations less than 50 km of the presumed core of the invasion range (Dumpert & Platen, 1985). Individuals in these areas were present for at least 45 years, during which local predators might have adapted to these novel prey. Indeed, a meta-analysis by Hawkes (2007) found that invasive plant species may accumulate novel enemies over time.

| D ISCUSS I ON
Additionally, another meta-analysis by Chun et al. (2010) showed that invasive plant species suffered relatively less damage than native species studied in the fields compared with greenhouses.
Hence, some invasive species may dominate in the fields where natural enemies do not recognize them as a suitable food source, but these enemies would feed on them in enclosed conditions (e.g., Siemann & Rogers, 2003;Lankau et al., 2004;Siemann et al., 2006; but see Carpenter & Cappuccino, 2005). Consequently, ecological and evolutionary processes that drive invasions might change over time (Hawkes, 2007) and different phases of the invasion process (Drake, 2003;Roy et al., 2011), whereby local predators may increasingly recognize invasive species as potential prey over time.
However, most studies on the enemy release hypothesis are fo- Indeed, a recent meta-analysis showed that spiders' total biomass across 54 North American grasslands failed to increase with total invertebrate biomass (Welti et al., 2020), indicating the potential control by their own predators (Sanders & Platner, 2007 also found no evidence for the role of soil disturbance (Narimanov et al., 2021) or higher competitive ability toward local sympatric species (Eichenberger et al., 2009). Therefore, other potential mechanisms behind their rapid spread and successful establishment, notably their high reproduction, remain to be investigated.

ACK N OWLED G M ENTS
The authors thank Linda Eberhardt for support with materials; Linda Eberhardt, Elena Hommel, and Anja Weiler for rearing colonies of springtails; and two anonymous reviewers for helpful comments on an earlier draft.

CO N FLI C T O F I NTE R E S T
None declared.

DATA AVA I L A B I L I T Y S TAT E M E N T
The dataset generated and analyzed during this study is available in