In today's globalized world, the introduction of exotic species in natural habitats represents a major conservation challenge (Sala et al. 2000). Once a species arrives to a new habitat, multiple biotic and abiotic factors determine whether it thrives and becomes part of the local community (Pearson et al. 2018; Weiher 1998). Among biotic filters, post-dispersal seed predation is an important demographic bottleneck in the life cycle of plants (Crawley 2000), which can strongly affect the capability of exotic species to become part of the recipient community. If generalist granivores preferentially consume exotic seeds, they will impose a biotic barrier to the invasion process (Maron et al. 2012; Nuñez, Simberloff, and Relva 2008; Pearson et al. 2014). In contrast, if exotic seeds are systematically avoided, they will have a recruitment advantage over natives (Connolly et al. 2014; Lucero and Callaway 2018; Pearson, Callaway, and Maron 2011). Therefore, understanding the feeding preferences of generalist granivores will improve our predictions about the potential of exotic species to invade new areas (Larios et al. 2017).
Among granivores, rodents are widespread generalist consumers whose foraging preferences can modulate the recruitment of exotic species (Connolly et al. 2014; Muschetto et al. 2015; Pearson et al. 2011). Even though factors driving rodent foraging choices can be multifaceted (e.g., coat harness, nutrient content, phenolic compounds (Blate et al. 1998; Gong, Tang, and Wang 2015; Sidhu and Datta 2015), seed size has been pointed out as the most important modulator of their eating preferences (Dylewski et al. 2020). In general, seed size effects are not unequivocal but depend on the range of seed sizes present in the community (e.g., positive effects of size in small-seeded grasslands, while negative in large-seeded forests (Radtke 2011). Such context-dependency may explain the contrasting results found in the literature with respect to the role of rodents in biological invasions, with large exotic seeds having high consumption rates (Nuñez, Simberloff, and Relva 2008) as well as escaping predation (Pearson, Callaway, and Maron 2011). It is expected that exotic seed with contrasting characteristics to those of natives will be perceived as a novel food item and they will have differential consumption rates (Pearson et al. 2014; Pearson et al. 2018; Pearson, Callaway, and Maron 2011). Also, whether exotic seeds are over or under-predated by granivorous, will depend on the sign of seed size effects in the community and the size of exotic seeds with respect to natives.
In addition to species-specific characteristics like seed size, granivores may optimize their foraging by consuming the most abundant species in the seed pool (Thompson, Brown, and Spencer 1991; Larios, Pearson, and Maron 2017). Such positive density-dependent foraging has been mainly tested at local scales and is explained by a higher probability of rodents to encounter and be attracted to areas with abundant seeds (e.g. Baraibar et al. 2012; Batisteli et al. 2020; Myster 2003; Wang 2020). In contrast, the effects of species relative abundance at landscape and community scales have been largely overlooked in spite that it can affect recruitment success at early stages of the invasion process. Before exotic species become invasive, their propagule pressure tends to be low relative to natives (Křivánek, Pyšek, and Jarošík 2006; Pyšek et al. 2015; Simberloff 2009), and hence, preferential consumption of abundant seeds may facilitate exotic species overpassing the post-dispersal predation barrier. Therefore, evaluating density-dependent seed predation at landscape scales will provide new insights about the potential advantages of being rare at initial stages of the invasion process.
Granivory, as any biotic interaction, can be modulated by the ecological setting in which plant-animal encounters occur. In particular, the structure of the vegetation can affect post-dispersal predation rates leading to high spatial variability in the community of recruits (Germain et al. 2013), including the establishment of exotic species (as observed in Muschetto et al. 2022). Lack of antipredatory cover (e.g., shrubs) can reduce the activity of rodents (García, Zamora, and Amico 2011; Kollmann and Buschor 2003) limiting the probability of seed encounter (Hulme 1994; Wang 2020). In addition, to avoid exposure to predators, rodents may reduce the time devoted to seed selection in open areas (Morán-López et al. 2018) becoming opportunistic foragers (Perea et al. 2011). Consequently, the probability of predation may not depend on seed traits, equalizing seed consumption rates across species. In the context of biological invasions, such shifts in rodent activity and behavior may turn open areas into invasion foci, where exotic species have a higher probability of escaping predation and recruiting (Pearson et al. 2014).
In this work, we estimated the combined effect of species-specific seed size, landscape-scale relative abundances, and shrub cover on predation rates of native and non-native species in a Patagonian temperate forest. In our study area, two species with contrasting seed sizes, Rubus idaeus and Prunus cerasus, have established a self-sustained population (i.e., naturalized). Seeds of Rubus ideaus are of similar size to those of natives, whereas seeds of Prunus cerasus are four times larger than the largest native seed (Fig. 1). Seeds of both species are relatively rare in the community, though during the fruiting season Rubus seeds are 1.5 times more abundant than those of Prunus (1.63% vs 0.93%). Finally, tree falls create forest gaps where the availability of antipredatory cover is low (García, Zamora, and Amico 2011). We expected a positive effect of seed size on predation rates given that seeds of the recipient community are small (< 5 mg) (Online Resource 1) (Dylewski et al. 2020). Consequently, rodents may preferentially consume large Prunus cerasus seeds, whereas Rubus idaeus seeds will be consumed at similar rates to natives. In addition, we expected rarity to confer an advantage to exotic species, since a positive density-dependent response to seed abundance has been previously observed in the area (García, Zamora, and Amico 2011). Finally, we expected that the biotic filter of seed predation will be attenuated in areas of low shrub cover due to decreased rodent activity and a less selective foraging (Germain et al. 2013).