Discussion

Total biomass of invasive plant species tended to higher than that of co-occurring native plant species, and the growth rate over time was also significantly higher than that of co-occurring native plant species regardless of the presence of soil microbes (Fig. 2A & Fig. S1). It is likely that invasive plants generally have a higher growth rate than native plants (van Kleunen et al. 2010), the current test invasive plant species also exhibited the trait as invasive plants grow larger than native plant species over time.
Soil microbes tended to had a slight positive effect on total biomass of the native plant species over time, while the effect of soil microbes on invasive plant species as their total biomass tended to change from promotion to inhibition over time (Fig. 2A). Consistent with our predictions, these results suggest that soil microbes affected the growth of invasive plant species and native plant species over time, and presence of soil microbes inhibited the growth of invasive plant species over time. This founding supports a prediction of declining enemy release over time that invasive plants accumulated more microbial natural enemies, and the increase in enemies (e.g., pathogen) may inhibit the growth of invasive plants themselves (Callaway et al. 2013, Diez et al. 2010, Mitchell et al. 2010, Stricker et al. 2016). Another possible explanation is that invasive plant responses to mutualisms weaken over time (Seifert et al. 2009), thus intensifies the inhibitory effect on the growth of invasive plants. In addition, the reduction of soil nutrients over time can also alter plant-soil feedback (Bennett et al. 2019), especially the rapid resource acquisition of invasive plants (Dukes et al. 1999), it may further amplify the negative feedback effects caused by pathogen accumulation or reduced mutualists dependence.
Local adaptations may be responsible for positive feedback in native plant species. Native plant species may be able to quickly adapt to the local soil microbial community (Kawecki et al. 2004). In generally, native plants tend to be conservative, have stronger tolerance to antagonists, and enhance adaptability to the local environment. Our research results showed that soil microbes always promote native plant species, which may be due to local adaptation in short-term, because it is found that this promotion effect weakens over time (first to second: 7.3% to 2.7%; Fig. 2A). We speculate that the promotion effect may also become inhibition in the future for a longer period, but the current results cannot prove it. Although the growth performance of invasive plant species was found to be different in individual analysis of species in each genus, and only the three-factor interaction was found in Paspalum (Fig. 2a-2), we speculated that temporal variations exist in the direction and strength of interactions between different plant species and soil microbes. Therefore, future studies may text the growth of different species over continuous time, to determine whether the strength and direction of invasive plant-to-soil feedback emerges with a general pattern.
Harvest time significantly decreased proportional allocation of biomass on roots (Fig. 2B), this result may be dominated by native plant species. We found that native plant species allocated proportionally biomass to their roots decreased over time, while an opposite pattern was observed for invasive plant species that tended to increase over time (Fig. 2B). It is consistent with plants growth responses, with invasive plant species allocated more root biomass to resource acquisition on growth. This is not surprising as more roots can help plants obtain more nutrients, and plants will preferentially allocate biomass to roots that can improve resource acquisition (Casper et al. 1997, Poorter et al. 2000). Although we found no significant three-way interaction between the invasion status, harvest time and soil treatments on root mass fraction (Fig. 2B), but the three-way interaction was observed in Paspalum (Fig. b-2) andSolidago (Fig. b-3). While the two genera did not show a same pattern, As guessed above, it is due to that the interaction between plant species and soil microbes is strongly related to temporal variations. In addition, we found that the presence of soil microbes significantly reduced the proportion of root biomass allocation, which indicates that soil microbes contain many pathogens, because soil-borne pathogens damage roots (Packer et al. 2003, Reinhart et al. 2010).