Statistical Analyses
All statistical analyses were performed using R Statistical Software (version 4.2.1, R Core Team 2022). Linear mixed models (LMM) were first used to examine the effects of rarity, phylogenetic relatedness of conditioned soil (soil origin), N enrichment, and fungicide application on total plant biomass, ECM colonization, and DSE colonization (“lmer” function in “lme4” package, R) (S1 Table ). The LMM yielded an interactive effect between rarity and soil origin, however, there was no main effect of fungicide nor interactive effects between rarity/soil origin and other fixed effects on total biomass. On the other hand, fungicide demonstrated a main effect on ECM colonization and DSE colonization. To account for the effects of N enrichment (low vs. high) and fungicide (live cultures vs. fungicide) application on total biomass, ECM colonization, and DSE colonization, N enrichment was utilized as a fixed effect and fungicide application was accounted for in the error structure of all linear mixed models as a random effect. Specifically, to address Hypothesis 1 and 2 that total plant biomass, ECM colonization, and DSE colonization vary based on the rarity of plant host species and the phylogenetic origins of conditioned soil inoculum under different levels of N enrichment, linear mixed models were used to determine the singular and interactive effects of rarity, soil origin, and N enrichment on the total biomass, ECM colonization, DSE colonization of eucalypts accounting for a blocked design in the error structure of the model. Tukey HSD post-hoc analyses were completed for all significant results (“ghlt” and “cld” functions in “multcomp” package, R).
Additionally, standardized plant-soil feedback mean effect sizes (Hedges’ g) were calculated for the total biomass and percent ECM colonization of plant species grown in conspecific versus same lineage soils, and conspecific versus opposite lineage soils (i.e., the soils include rare and common species within and across lineages) on the basis of the 2nd generation of plants that varied in level of rarity, described above. Effect sizes were calculated using the Hedges’ g statistic as it accounts and corrects for variance that may be introduced by unequal or small sample sizes (Hassan et al. 2022). Positive values of Hedges’ g indicated positive plant-soil feedbacks in which the total biomass or percent ECM colonization was greater in soils conditioned by similar or opposite lineages compared to conspecific soils. These types of positive feedbacks in biomass and mycorrhizal colonization mediated by phylogenetically similar soil conditions represent a home-field advantage in which phylogenetically similar plant species readily share microbial mutualists due to a shared evolutionary history and symbiosis of host-associated microbial communities (Anacker et al. 2014; Fitzpatrick et al. 2018; Kohl 2020; Kazarina et al. 2023). On the other hand, negative values of Hedges’ g indicate negative plant-soil feedbacks in which the total biomass or percent ECM colonization was lower in soils conditioned by similar or opposite lineages compared to conspecific soils. These types of negative feedbacks in biomass and mycorrhizal colonization mediated by phylogenetically similar or dissimilar soil conditions are indicative of a disruptive effect on plant growth and inform the maintenance of trait variation within plant communities. Hedges’ g values that were not significantly different than 0 were representative of a neutral or lack of plant-soil feedback. Therefore, the Hedges’ g values associated with each comparison along the rarity gradient represented a measure of both the strength and direction of microbially mediated plant-soil feedbacks. The strength of plant-soil feedbacks was assessed by the magnitude of the effect size, while the direction of the feedbacks was determined by the sign of the effect size. Pearson product-moment correlations between the effect sizes of total biomass and ECM colonization along a rarity gradient were also analyzed to further understand the significance and direction of feedbacks.

Results

In support of Hypothesis 1 that total plant biomass, ECM colonization, and DSE colonization are related to the rarity of plant host, rare species displayed significantly lower biomass, but higher rates of ECM colonization than common counterparts when grown in a common garden environment across all soil conditions. (Table 1 ;Figure 1 ). For example, the rarest species had 62% more ECM colonization while maintaining 88% less total plant biomass than common counterparts (Figure 1). The rarest species also demonstrated 64% greater DSE colonization in soil than common species, however, this difference was not significant at α=0.05. Further, the relationship between total plant biomass and ECM colonization was significantly dependent on soil N enrichment (Table 1). Eucalypt species across all rarity levels grown in high N conditions demonstrated a 37 % increase in biomass but 11 % decrease in ECM colonization than species under low N conditions (Figure 2 ). Consequently, high N conditions significantly facilitated higher levels of plant biomass but lower ECM colonization, while low N conditions significantly supported lower levels of plant biomass but higher ECM colonization. Additionally, rarity and N treatment demonstrated a significant interaction to affect the ECM colonization of Tasmanian eucalypts (Table 1; Figure 2). The disparity between rare versus common biomass and ECM colonization was more prevalent under high N enrichment, as demonstrated by a 166% increase in biomass and a 64% decrease in ECM colonization from rare to common species under high N conditions, compared to a 149% increase in biomass and a 27% decrease in ECM colonization from rare to common species under low N conditions.
In support of Hypothesis 2 that total plant biomass, ECM colonization, and DSE colonization are related to the phylogenetic origin of conditioned soil, there was a significant interaction between plant rarity and soil origin; rare species maintained significantly lower biomass but higher rates of ECM colonization in same lineage and opposite lineage soils compared to conspecific soils (Table 1;Figure 3 ). For example, the rarest species, grown in soil conditioned by phylogenetically similar species demonstrated a 114% increase in ECM colonization, but a 94 % decrease in total biomass compared to conspecific conditioned soils. These patterns were also consistent in species grown in phylogenetically distant soil which demonstrated a 96 % increase in ECM colonization, but a 71 % decrease in total biomass compared to conspecific conditioned soils. On the other hand, common species displayed similar patterns, however, to a much lesser extent; ECM colonization increased by 26%-78% and total biomass decreased by 25%-52% in conspecific relative to distantly related soil inoculum and conspecific to closely related soil inoculum, respectively. DSE colonization did not vary significantly by phylogenetic origin of conditioned soil across all rarity levels. These results from a common garden suggest that rare species may maintain distinct and critical tradeoffs between aboveground performance traits and belowground associations with fungal mutualists not prevalent in common species (i.e., there is reduced performance but increased reliance on ECM in any different soil than that conditioned by the same species in rare eucalypts).
Consistent with Hypothesis 3, that the divergence in plant total biomass in conspecific versus soils conditioned by other species is driven by rates of ECM colonization, biomass exhibited strong negative plant-soil feedbacks in conditioned soil relative to conspecific soil, while rates of ECM colonization demonstrated simultaneous strong positive plant-soil feedbacks (Figure 4 ). These results suggest that species grown in soil conditioned by other species maintain lower biomass but stronger associations with mycorrhizal mutualists when compared to the same species grown in monoculture. Furthermore, the effect size of ECM colonization directly affected the effect size of total biomass in phylogenetically distant conditions with a significant negative correlation (r^2 = -0.83) to show that performance traits, such as biomass, may be a function of ECM colonization acting differently among species varying in rarity (Figure 5 & Table 2 ). Further, the strength of the relationship between belowground mycorrhizal feedbacks and aboveground plant traits was significantly affected by the phylogenetic relatedness of conditioned soil. While both conspecific to same lineage and conspecific to opposite lineage treatments demonstrated similar patterns in the relationship between total biomass and rates of ECM colonization, the differences in feedbacks in conspecific to same lineage treatments displayed a moderate negative relationship (r^2 = -0.58). Differences underlying the significance of conspecific to same lineage and conspecific to opposite lineage feedbacks may be due inadequate sample size. These results suggest that differences in ECM colonization not only drive aboveground feedbacks differently in rare versus common species, but also more strongly in distantly related species. Taken together, these results suggest that microbially mediated belowground feedbacks function to drive aboveground feedbacks in performance traits critical for species survival. Furthermore, if the continued survival of rare species is dependent on the maintenance of low biomass via belowground microbial associations, then the legacy effects of soil conditioned by species varying in rarity and phylogenetic relatedness are expected to exert strong control over future plant community phenology, composition, and biodiversity at the population level.

Discussion

Species rarity is often considered an unfavorable ecological response to environmental factors (Wamelink et al. 2014; Irl et al. 2017). Recent studies, however, have demonstrated that ecological and evolutionary factors interact to drive divergent patterns in aboveground traits between rare and common species (Boyd et al. 2022; Nytko et al. in review ). Much less consideration has been given to potential belowground mechanisms driving the aboveground divergence of traits and subsequent plant-soil feedbacks in rare species. This omission is important as species do not live in isolation; community context and soil legacy affects are meaningful and likely have important feedbacks that can reinforce rarity over time. Our results show that 1) the total biomass and ECM colonization of a species is dependent on the rarity of the plant host; 2) the total biomass and ECM colonization of a species is also affected by the phylogenetic relatedness and N enrichment of conditioned soil; and 3) differences in belowground ECM colonization drive aboveground trait divergence and larger plant-soil feedbacks in rare versus common species. Taken together, the results suggest that rare plants may more heavily utilize mycorrhizal associations in phylogenetically distant communities under low N conditions, than common species, even while maintaining lower biomass levels for continued persistence across the landscape. Overall, rare plant species, while having reduced biomass aboveground, may maintain rich belowground mycorrhizal, rather than non-mycorrhizal, fungal communities conducive to nutrient acquisition and high potential for adaptation under climatic change. Results such as these have not commonly been considered for rare species.
Plant performance is often measured using a variety of functional traits such as life-form, dispersal, seed biology, and phenotypic response to environmental conditions (Pywell et al. 2003, Bragion et al. 2018; Hanisch et al. 2020). However, without an understanding of the mechanisms underlying functional trait expression, predictions of trait response to climate and associated changes in ecosystem function are limited. For example, our results demonstrate that rare species overall maintain lower biomass than common species across all soil conditions. Under current measures of plant species’ success, lower productivity is often considered ecologically unfavorable due to associations between lower biomass and decreased fecundity (Poorter et al. 2015; Younginger et al. 2017). However, our results also indicate that the aboveground trait values of rare species may not be representative of an unsuccessful survival strategy, as is often seen in common species. Instead, lower biomass in rare species may reflect a unique tradeoff in which more resources are allocated to belowground mycorrhizal symbionts for nutrient uptake and pathogen protection that allow for continued persistence (Poot & Lambers 2003; Bothe et al. 2010; Soudzilovskaia et al. 2019; Tedersoo & Bahram 2019). Furthermore, the trade-off between ECM fungal colonization and aboveground biomass is dependent on the phylogenetic relatedness of conditioned soil in which plants were grown and the rarity level of plant host, such that the rarest species experience the strongest negative feedbacks in growth but strongest increase in ECM colonization when grown in soil conditioned by phylogenetically distant species. These conclusions support and expand upon the findings of Kempel et al. (2018) and Jiang et al. (2022) to demonstrate the microbial mechanisms underlying plant-soil feedbacks in plant communities varying in rarity and phylogenetic distance.
While these patterns may be mediated by the unique functional traits of rare species varying in phylogenetic relatedness, an alternative hypothesis may suggest that the characteristic low biomass of rare species is a function of ECM parasitism in low N environments. While these results are consistent with parasitic patterns in belowground fungi, as well as the functional equilibrium model of resource allocation under N enrichment (Johnson et al. 2008), it is also unusual to observe parasitism of mycorrhizal fungi in low N environments in which mutualisms are necessary for nutrient acquisition (Johnson 1993; Johnson et al. 2008). Consequently, the biomass of rare plant species may be representative of: 1) N-limitation; 2) allocation to belowground resources for limiting micronutrients; 3) the maintenance of valuable functional traits rather than biomass production in nutrient-lacking environments. Although the effects of soil nutrient conditions on productivity in rare versus common plant species need to be studied further, these results suggest that the ability of a rare species to survive and persist in expanding range and niche space may be microbially mediated, particularly via previous plant communities. Specifically, rare plant species may benefit geographically, competitively, and functionally from mycorrhizal mediated soil legacy effects, to facilitate the persistence and potential expansion of rare species in phylogenetically distant, and to a lesser extent phylogenetically similar, communities rather than common species under climate change.
Predicting the future spatial and temporal persistence of rare species is key to conservation and management, as well as maintaining ecological function (Kunin & Gaston 1997). Species’ rarity can be predicted by prominent negative plant-soil feedbacks (Ke et al. 2015; Klironomos 2002), however, predictions of rare species persistence are limited without a mechanistic understanding of these plant-soil feedbacks. Studies examining the mechanistic causes of rarity are often ecological in focus, however, recent research has demonstrated that the phylogenetic origin of plant species and the microbial community of previously conditioned soil can interact to affect the strength and direction of mycorrhizal mediated feedbacks (Segnitz et al. 2020). For example, Woolbright et al. (2014) suggested that novel interactions with ECM fungi may influence the ability of relict species to migrate, persist, or outperform common counterparts, while maintaining lower levels of genetic variability, in atypical habitats (Chung et al. 2015). Our results not only confirm that rare species on average suffer stronger negative plant-soil feedbacks on biomass when grown in conspecific soils, but also show that the strength and direction of plant-soil feedbacks in rare species are mediated by positive feedbacks in ECM colonization in both closely and distantly related plant-conditioned soils. These results suggest that conspecific-conditioned soils may be lacking adequate levels of mycorrhizal mutualists, while incurring high levels of plant pathogens (Hannula et al. 2021). Consequently, rare species productivity is inextricably tied to belowground mycorrhizal colonization, making rare species more susceptible (both positively and negatively) to soil conditioning effects such as legacy and priority effects as well as above- and belowground species interactions (Nytko et al. in review ). Therefore, long-lasting microbially mediated legacy effects of previous plant communities may drive patterns of rare species abundance, as demonstrated by low abundant species in Anacker et al. (2014), Maron et al. (2016), and Yan et al. (2022), as well as patterns of rare species productivity and composition in successional communities.
Biotic interactions shape species’ distributions, functional traits, and rarity (Cosentino et al. 2023; Kempel et al. 2020; Nytko et al. in review ; Wisz et al. 2013); however, biotic interactions are commonly underrepresented in conservation, leading to errors in the estimation of rare species occurrence and future persistence (Flores-Tolentino et al. 2020). Although the majority of literature regarding the causes and consequences of rarity is dedicated to the response of species to abiotic factors, Woolbright et al. (2014) demonstrated the critical importance of biotic interactions in understanding both the geographic occurrence and genetic underpinnings of isolated relict species. Our results expand on the hypotheses proposed in Woolbright et al. (2014), to demonstrate the critical importance of above-and belowground interactions in understanding rarity as well as determining functional trait expression (Kempel et al. 2020). These results suggest that the expression of phenotypes among populations within the same species are the consequence of previous plant communities and current feedbacks with belowground symbionts. Moreover, biotic interactions may drive patterns of geographic mosaics in rare species by shaping the direction of co-evolution to establish differences in trade-off strength and subsequent trait expression not seen in common species. Consequently, geographically based differences in the availability and abundance of mycorrhizal symbionts may result in population-related variation in the expression of traits, long-term species persistence, and conservation priority. As such, the phenotypes of rare species that we consider to be a conservation priority may be a consequence of microbially mediated plant-soil feedbacks.
Conclusions & Implications : The majority of plant species are rare in any given plant community. Furthermore, the conservation of rare populations are often characterized by geographic and genetic parameters, such that a rare population is one of small size, located in an uncharacteristically atypical area, with limited genetic diversity (Rabinowitz 1981; Cole 2003) without consideration of the community context or past evolutionary history of that species. Rare species are rare for a variety of reasons, but some may maintain lower levels of biomass and higher rates of ECM colonization to persist. If the microbially mediated plant-soil feedbacks seen in rare species are representative of a unique trade-off between allocation to ECM colonization versus aboveground biomass when growing in intermediately related plant-conditioned soils, then there are likely to be significant implications for conservation practices. Although abiotic factors are important proponents of functional trait expression (Lavorel et al. 2011; Caruso et al. 2020), the long-lasting microbial impacts of previous plant communities, especially ones that may not strongly alter soil communities in positive or negative ways for the next generation of plants, may exert more control over the phenotypic appearance and persistence of rare species than originally hypothesized. As legacy effects become increasingly pronounced under climate change, accounting for above- and belowground biotic interactions will become increasingly important in spatial distribution modeling (SDM) and predictions of rare species occurrence in changing communities. Additionally, if species rarity is dependent on microbial counterparts for the successful maintenance of lower biomass, rare species management may be improved through the consideration of belowground colonization via mycorrhizal mutualists at the population level. These greenhouse results suggest that the aboveground traits of rare species are not indicative of an unsuccessful plant strategy, but rather a consequence of previous plant conditioning effects as well as interactions with current fungal mutualists that vary by previous plant community inhabitants that need to be field-tested. Accordingly, the low abundance and biomass characteristic of rare species may disguise a rich belowground world of beneficial mutualists, worthy of further investigation into the contribution of belowground feedbacks to aboveground functional traits in rare species around the world.