1. Understanding mechanisms underlying species distributions is a central theme in ecology. This study identifies factors driving spatial variation in arbuscular mycorrhizal fungi (AMF). We conducted two experiments to test whether heterogeneity in AMF colonization of alpine perennial plants across a willow-meadow ecotone is due to variation in (i) above-ground competition with willows for light (experiment 1), (ii) below-ground interactions with willows and their ectomycorrhizal fungi (EMF; experiment 1) or (iii) leaf litter deposition (experiment 2).
2. In experiment 1, we tested the above-ground interactions hypothesis by covering open meadow plots with 80% shade cloth to simulate willow shading (S). To test the below-ground interactions hypothesis, we transplanted ectomycorrhizal (MW) and nonmycorrhizal willows (NW) into the open meadow. AMF colonization of herbaceous plants in the S, MW and NW treatments was compared to colonization of plants growing in unmanipulated open meadow (OC) and willow understorey (WC) control plots. In experiment 2, we tested the leaf litter hypothesis by manipulating leaf litter deposition in open meadow and willow understorey plots. AMF and EMF colonization was compared in plots with and without leaf litter.
3. In experiment 1, AMF colonization was reduced in MW and WC plots compared to the other three treatments, suggesting that below-ground interactions with EMF suppressed AMF colonization of herbaceous hosts. In experiment 2, the presence of leaf litter increased EMF colonization in the open meadow and reduced AMF colonization in both open meadow and willow understorey habitats, suggesting that willow-derived leaf litter indirectly affected AMF colonization by promoting EMF colonization.
4.Synthesis. Our results indicate that willows indirectly reduce AMF colonization of neighbouring herbaceous plants via feedbacks with leaf litter and EMF. These willow-mediated effects could alter the distribution of mycorrhizal fungi in alpine communities, which could in turn impact the fitness and distribution of closely associated host species. Ultimately, this study demonstrates the potential for below-ground interactions to drive variation in species associations across ecotonal boundaries.
Mycorrhizal associations are widespread, ecologically important interactions between plants and root-colonizing fungi that can impact individual plant growth and competitive ability as well as overall biodiversity and productivity (van der Heijden et al. 1998; Wilson et al. 2001). Mycorrhizal effects often vary spatially and temporally, highlighting the context-dependent nature of these associations (Hoeksema et al. 2010). Identifying factors that influence mycorrhizal associations may inform questions about their role in shaping plant communities across heterogeneous landscapes.
Alpine ecosystems provide excellent opportunities for studying context-dependency in mycorrhizal associations due to the prevalence of nutrient limitation coupled with spatial variation in mycorrhizal fungi. Soil resources are heterogeneously distributed and difficult to capture from cold alpine soils, thereby imposing nutrient limitation on alpine plants (Karlsson & Nordell 1996). Given their positive effect on plant stress tolerance and resource uptake (Smith & Read 1997), mycorrhizal associations are likely important in alpine ecosystems (Scherff, Galen & Stanton 1994; Theodose & Bowman 1997); yet, surveys show that mycorrhizal status varies greatly in these communities (Read & Haselwandter 1981; Cripps & Eddington 2005). At our study site in the central Rocky Mountains, a broad habitat mosaic formed by willow and conifer outcroppings in tundra meadow vegetation delineates the transition zone between subalpine forest and alpine tundra habitats. In this zone, known as the krummholz because of the presence of twisted and stunted trees, mycorrhizal associations vary across the willow-meadow ecotone. Specifically, colonization by arbuscular mycorrhizal fungi (AMF) in several common tundra forbs is greater in plants growing in the open meadow than in the willow understorey (Becklin & Galen 2009). If AMF are important to nutrient uptake or stress tolerance at this site, then suppression of AMF colonization in understorey habitats could impact the growth of tundra plants, limiting their downward distribution into subalpine communities.
Alpine willows (Salix sp.) could affect AMF colonization by competing for resources that influence the cost-to-benefit ratio of mycorrhizal associations. AMF are obligate biotrophs that rely entirely on their host plant for carbon; 4–20% of a host plant’s total carbon budget may be allocated to mycorrhizal symbionts (Johnson, Graham & Smith 1997). Reduced irradiance may limit photosynthesis and the capacity of plants to support extensive mycorrhizal networks (Tester et al. 1986; Son & Smith 1988). As willow canopies dramatically reduce the amount of light available to understorey plants (Totland & Esaete 2002; Dona & Galen 2006), carbon limitation may negatively affect AMF colonization of herbaceous hosts (above-ground interactions hypothesis).
Alternatively, willows could affect AMF colonization by competing for below-ground resources that are required for both plant and fungal growth. For example, as leaf nitrogen content is tightly linked to photosynthesis, competition for nitrogen may restrict the ability of plants to supply their fungal partners with carbon (Johnson 2010). Resource competition may also limit the amount of nutrients directly available for fungal growth, thereby affecting colonization regardless of carbon supply (Treseder & Allen 2002). Willows are dominant species in krummholz plant communities and their extensive root system may usurp soil resources from smaller herbaceous competitors. Willows also compete indirectly for below-ground resources via their ectomycorrhizal fungi (EMF); these fungi produce extensive hyphal networks, interact closely with co-occurring saprotrophic fungi, and have a greater capacity to break down organic substrates than AMF (Lindahl et al. 1999; Read & Perez-Moreno 2003). Consequently, EMF may outcompete their AMF counterparts for resources such as organic forms of nitrogen. If competition for below-ground resources restricts either fungal growth or photosynthesis, then willows and their EMF partners may suppress AMF colonization in understorey habitats where the densities of willow roots and EMF are high (below-ground interactions hypothesis).
Allelopathic compounds released by willows could affect AMF colonization independent of competition for resources. Depending on chemical composition and concentration, plant secondary compounds can increase or decrease fungal growth (Fries et al. 1997; Piotrowski, Morford & Rillig 2008). Willows produce a number of secondary compounds that are released as root exudates or during leaf litter decomposition (Schmidt, Lipson & Raab 2000; Hansen et al. 2006). EMF can also produce secondary compounds that inhibit the growth of microbial competitors (Garrido et al. 1982). Both willow- and EMF-derived secondary compounds are more likely to accumulate in the understorey where willow roots, EMF and leaf litter are abundant. If these secondary compounds negatively affect the growth of AMF or their hosts, then allelopathy could explain the observed pattern in AMF colonization in krummholz habitats. Because allelopathy and below-ground competition are difficult to distinguish in the field, we view both mechanisms as components of the below-ground interactions hypothesis.
Finally, leaf litter deposition is another indirect mechanism by which willows could influence AMF colonization of herbaceous neighbours. Leaf litter is a potential source of carbon and nitrogen for soil organisms; however, fungi differ in their ability to decompose leaf litter and access these resources (Read & Perez-Moreno 2003). These functional differences can lead to positive feedbacks between dominant host plants and their mycorrhizal partners if those fungi are better able to decompose host-derived leaf litter (Wurzburger & Hendrick 2009). Alpine willows are deciduous and produce a substantial amount of leaf litter; on average, 0.6 g cm−2 of leaf litter accumulates in the willow understorey at our study site (K. Becklin, unpubl. data). As EMF often interact with saprotrophic fungi (Lindahl et al. 1999) and decomposition of organic substrates is thought to be greater for EMF compared to AMF (Read & Perez-Moreno 2003), willow-derived leaf litter may increase the activity of EMF to the detriment of co-occurring AMF and their host plants (leaf litter hypothesis).
Historically, plant ecologists focused on factors that limit the upward migration of woody plants into alpine habitats (Körner 2003); in this study, we examined potential mechanisms by which woody plants may limit the downward distribution of herbaceous tundra plants and their fungal partners. Through two field experiments, we asked whether (i) above-ground interactions (light competition), (ii) below-ground interactions (resource competition or allelopathy) or (iii) leaf litter deposition drive AMF colonization across the willow-meadow ecotone. In this study, we focus on AMF colonization at the community level, recognizing that there is substantial variation in colonization and mycorrhizal effects among host species. Assessing mycorrhizal effects on individuals of slow-growing, long-lived alpine plant species would require long-term monitoring, which is beyond the scope of this study. Our results indicate that willows indirectly suppress AMF colonization in understorey communities through feedbacks with EMF and leaf litter deposition. We discuss the implications of these results for the hypothesis that the interactions between co-occurring plant–fungal mutualist guilds can influence species distributions across ecotonal boundaries.
Materials and methods
This study was carried out at three elevational sites spanning the krummholz transition zone between subalpine forests and alpine tundra communities on Pennsylvania Mountain (Park County, CO, USA; 3682, 3626, and 3588 m a.s.l.). At this location, a mosaic of open meadow and willow understorey microsites dominates the krummholz region. Alpine willows (Salix glauca and Salix brachycarpa) play an important role in these ecosystems, insulating herbaceous plants during the winter and competing for light, water and nutrients during the growing season (Totland & Esaete 2002; Dona & Galen 2006, 2007). As a result, willows can greatly impact the surrounding plant community and potentially the community of organisms that associate with alpine perennials. This study was conducted across the willow-meadow ecotone in nutrient-limited dry alpine meadows where mycorrhizal associations are predicted to be particularly important (Theodose & Bowman 1997).
In krummholz regions of the central Rocky Mountains, S. glauca is the primary ectomycorrhizal (EM) host, although stunted conifer trees (Abies lasiocarpa, Picea engelmannii and Pinus aristata) also occur. In other systems, willows are colonized by both EMF and AMF (Wang & Qiu 2006); however, using staining techniques we have not seen evidence of AMF colonization of willows at this location in 4 years of study (K. Becklin, pers. obs.). A survey of EMF colonization across the willow-meadow ecotone indicates that, as expected, EMF are more abundant in the willow understorey where the density of willow roots is high (Becklin & Galen 2009). On average, EMF colonize 33% (range = 0–64%) of root tips in samples collected from the willow understorey compared to 2% (range = 0–33%) in samples collected from the open meadow.
Dry alpine meadows contain a diverse mixture of mycorrhizal and nonmycorrhizal grasses, sedges and forbs. Many of these herbaceous plant species occur in both open meadow and willow understorey habitats; however, plant density and diversity are generally higher in the open meadow (Becklin 2010). Arbuscular mycorrhizal (AM) hosts are common within the herbaceous community although other types of mycorrhizal fungi, including EMF, likely colonize some herbaceous alpine plants (Cripps & Eddington 2005; Wang & Qiu 2006). Surveys of four common herbaceous forbs, Taraxacum ceratophorum, Taraxacum officinale, Polemonium viscosum and Polemonium delicatum, indicate that AMF composition is similar across the willow-meadow ecotone (Becklin 2010), whereas AMF colonization is greater in the open meadow (Becklin & Galen 2009).
Effects of Above- and Below-Ground Interactions
To test whether above- or below-ground interactions drive spatial patterns in AMF colonization across the willow-meadow ecotone, we manipulated three aspects of the willow habitat (canopy cover, willow presence and EMF abundance) then monitored AMF and EMF colonization over time. This experiment was replicated in 2007–08 and 2008–09.
To isolate the below-ground effects of willows and EMF, we transplanted S. glauca cuttings into open meadow microsites then manipulated subsequent colonization by EMF. Willow cuttings were collected from source populations on Pennsylvania Mountain in 2002 (Dudley & Galen 2007), rooted in a sterile 1 : 3 mixture of sand and peat moss, and grown for 5–6years in the greenhouse (University of Missouri, Columbia, MO, USA). In early June of 2007 and 2008, we transplanted the willow cuttings back into the open meadow. At the time of transplanting, willows were 40–93 cm (average = 63 cm) tall with 3–6 (average = 3) branches, and 0–21% (average = 6%) of root tips were colonized by EMF. To control for differences in initial size, willows were transplanted in pairs of approximately equal size at least 3 m from each other (Fig. 1). Willow transplants were placed in holes twice the size of the rootball, and the extra space was filled with sterile peat moss. For one randomly selected willow within each pair, we lined the transplant hole with landscaping cloth (DeWitt Pro-5 Weed Barrier; DeWitt Company, Sikeston, MO, USA). This cloth served as a barrier minimizing contact with EMF present in the field soil, resulting in willow plots with a low abundance of EMF [nonmycorrhizal willow (NW); Fig. 3a]. The other willow within each pair was left open to colonization by naturally occurring fungi over the course of the experiment, resulting in plots with a high abundance of EMF [ectomycorrhizal willow (MW); Fig. 3a]. Compared to naturally occurring willows, the transplants had very small canopies ( >150 vs. <50 cm in diameter); thus, potential canopy effects were likely minimal in NW and MW treatments. After 2 years, the willow transplants were excavated. The root mass of NW and MW willow transplants was visually similar in size, indicating that the landscaping cloth did not restrict root growth of NW transplants. At that time, we also assessed final EMF colonization of each willow transplant to verify treatment efficacy and mycorrhizal status (see below for method of determining EMF colonization). To isolate shading effects independent of the presence of willows or EMF, we placed 80% shade cloth (Gempler’s, Janesville, WI, USA) over open meadow plots (S). We also established unmanipulated control plots in the open meadow (OC) and understorey of naturally occurring willows (WC). Individual plots were c. 0.25 m2 and there were 10 replicates per site for a total of 30 replicates per treatment. All treatments within a replicate were at least 3 m from each other and from the nearest naturally occurring willow (Fig. 1). Due to elk damage and transplant mortality, the total sample sizes for NW and MW treatments were reduced to 18 and 21, respectively.
To assess treatments effects on mycorrhizal fungal colonization, we collected soil cores (6 × 10 cm) in July of each year. Soil cores were collected from as close to the centre of each plot as possible and stored at 4 °C until processed. We isolated and washed the roots from each core; we then examined 300 root tips for EMF colonization using a dissecting microscope at 10× magnification. EMF root tips were identified based on shape, colour, branching pattern and the presence of a mantle. EMF colonization was calculated as the proportion of EMF root tips. After examining the roots for EMF, we subsampled the herbaceous roots from the soil core, avoiding woody roots that were likely from willow hosts. The herbaceous roots were cleared in 10% KOH at room temperature for 48 h, briefly acidified in 0.1 N HCl, stained with 0.05% trypan blue for 24 h and destained in lactoglycerol for 24 h (Phillips & Hayman 1970). Using a compound microscope at 400× magnification, we examined a total of 100 points along ten haphazardly selected root fragments for AMF colonization (McGonigle et al. 1990). AMF colonization was calculated as the proportion of herbaceous root length containing AMF structures (i.e. arbuscules, vesicles and coils).
As mycorrhizal fungal colonization depends on the presence of appropriate hosts, variation in plant community composition among plots could influence the results of this experiment. We used two approaches to account for variation in the herbaceous community among the five treatments. First, for each plot, we estimated the percentage cover by graminoid and forb host species in both years of the experiment. Second, we transplanted adults of four dominant alpine forbs at our study site (P. viscosum, P. delicatum, T. ceratophorum and T. officinale) into each plot, including the OC and WC. These species were examined in a previous survey of mycorrhizal associations on Pennsylvania Mountain, and results from that study indicate that AMF readily colonize all four plant species (Becklin & Galen 2009). The herbaceous transplants were collected from resident populations on Pennsylvania Mountain in early June of each year. Although these transplants only comprised a small portion of the total plot vegetation (average = 30%), transplanting these species into all five treatments decreased variation in plant community composition that could obscure treatment effects. By the end of the experiment, AMF colonized all herbaceous transplants (see Table S1 in the Supporting Information).
Abiotic conditions and resource availability can also impact mycorrhizal colonization. To evaluate whether abiotic conditions varied among treatments, we monitored air temperature, relative humidity and photosynthetic photon flux density (PPFD) in a subset of the plots at each site (N = 6 per treatment). Air temperature and relative humidity were measured using hobo H8 environmental data loggers (Onset Computer Corporation, Bourne, MA, USA). PPFD was measured using quantum light sensors connected to a CR-10 environmental data logger (Campbell Scientific, Inc, Logan, UT, USA). Data were collected between 09:00 and 15:00 on 13–18 July 2007. We also collected soil cores (2 × 10 cm) from a subset of plots in mid-July of each year after at least 3 days without rain to determine soil pH, percentage organic matter, extractable Bray I phosphorus concentration, extractable nitrate concentration, extractable ammonium concentration, exchangeable calcium concentration, exchangeable magnesium concentration, exchangeable potassium concentration and soil water content (N = 6 per treatment; Soil and Plant Testing Laboratory, University of Missouri, Columbia, MO, USA). Methods for the soil analyses are listed in Appendix S1 (see Supporting Information).
Effects of Leaf Litter Deposition
To test whether willow-derived leaf litter affects AMF colonization, we conducted a second experiment in 2008–09. Fifteen plots were randomly selected in willow understorey and open meadow habitats at each of the three elevational sites. Open meadow plots were at least 3 m from the nearest willow (Fig. 2). Each plot was divided into two circular subplots (20-cm diameter). In early June, we removed the leaf litter from one subplot in the willow understorey and added it to a subplot in the open meadow (Fig. 2). In August 2008, we collected a soil core (6 × 10 cm) from the centre of each subplot in seven replicates per site, and in August 2009, we collected soil cores from the remaining eight replicates per site. We used the same methods as in the first experiment to determine EMF and AMF colonization in each subplot. We also measured leaf litter biomass in each subplot after drying it at 60 °C for 48 h.
Variation in environmental conditions in the first experiment (soil pH, percentage organic matter, soil nutrient concentrations and soil water content) was analysed using multivariate analysis of variance (manova) with site (3682, 3626 and 3588 m a.s.l.), treatment and their interaction as fixed factors. A separate analysis was conducted for above-ground conditions (air temperature, relative humidity and PPFD) as these variables were only measured in S, OC and WC treatments. Differences in percentage plant cover over time were analysed using repeated measures analysis of variance (rm-anova) with site, treatment and their interaction as fixed factors. Treatment effects on colonization by EMF (proportion of colonized root tips) and AMF (proportion of colonized root length) over time were also assessed using rm-anova with site, treatment and their interaction as fixed factors. As arbuscules are indicative of functional AM associations, we conducted this analysis for arbuscular colonization as well as total AMF colonization. Planned contrasts were used to evaluate whether the presence of a canopy (S vs. OC), willow (NW vs. OC) or EMF (MW vs. NW) affected AMF colonization. Planned contrasts were also used to evaluate whether the effect of EMF on AMF colonization was similar in the experimental and control plots (MW vs. WC). The quantitative relationship between EMF and AMF colonization was evaluated using analysis of covariance (ancova) with treatment as a fixed factor. To evaluate the effect of soil conditions on EMF and AMF colonization, we first conducted a principle components analysis to transform the soil components (soil pH, percentage organic matter, soil nutrient concentrations and soil water content) into predictor variables. We then analysed the relationship between mycorrhizal colonization and each principle component (eigenvalue > 1) using ancova with treatment as a fixed factor.
In the second experiment, we evaluated the effect of leaf litter on EMF and AMF colonization using anova with year, site, habitat, leaf litter treatment and their interactions as fixed factors and replicate as a random factor. The quantitative relationship between EMF and AMF colonization in plots with and without leaf litter was evaluated using ancova with site, habitat, litter treatment and their interactions as fixed factors. The effect of leaf litter biomass on EMF and AMF colonization was also evaluated using ancova with site, habitat and their interaction as fixed factors.
All statistical analyses were conducted with the statistical program jmp 8.0 (SAS Institute Inc., Cary, NC, USA). Normality and homoscedasticity were improved by arcsine-square root transforming AMF and EMF colonization, percentage plant cover and soil water content. Soil pH, phosphorus concentration and nitrate concentration were log transformed. All other variables met the assumptions of normality and equal variances.
Effects of Above- and Below-Ground Interactions
Above-ground environmental conditions varied among treatments (manova: Ftrt = 3.81, d.f. = 6, P =0.02; Table 1). Air temperature was highest in OC plots compared to S and WC plots (planned contrast: F =11.77, d.f. = 1, P =0.01). PPFD was also higher in OC plots compared to S plots (planned contrast, F =15.16, d.f. = 1, P =0.004) and higher in S plots compared to WC plots (planned contrast, F =6.95, d.f. = 1, P =0.03). The shade cloth reduced PPFD by an average of 47% (±8% SE) while the unmanipulated willows reduced PPFD by an average of 79% (±5% SE). Relative humidity (Ftrt = 0.06, d.f. = 2, P =0.94), soil nutrient and water content (manova: Ftrt = 0.69, d.f. = 40, P =0.87), and percentage plant cover (rm-anova: Ftrt = 1.25, d.f. = 4, P =0.30) did not vary among treatments.
Table 1. Variation in environmental conditions among the five treatments: open meadow control (OC), shade cloth (S), nonmycorrhizal willow (NW), ectomycorrhizal willow (MW) and willow control (WC). Values represent the mean (SE)*
N = 30
N = 30
N = 18
N = 21
N = 30
*Letters indicate means that are significantly different at P <0.05 using Student’s t-tests.
†Variables were measured on two replicates per site (N = 6 per treatment).
‡Variables were measured for all plots in the experiment (N = 18–30 per treatment) and values represent the 2-year average.
PPFD, photosynthetic photon flux density; SWC, soil water content.
Air temperature (°C)†
Relative humidity (%)†
PPFD (μmol m−2 s−1)†
Organic matter (%)†
SWC (g g−1)†
Graminoid cover (%)‡
Forb cover (%)‡
EMF colonization of woody roots from the soil cores varied among treatments (rm-anova: Ftrt = 143.07, d.f. = 4, P = 0.0001; see Table S2 in the Supporting Information). As expected, EMF colonization was significantly higher in MW plots than NW plots (planned contrast: F =103.80, d.f. = 1, P =0.0001; Fig. 3a). However, EMF colonization was an additional 15% (±4% SE) higher in the WC plots (planned contrast: F =9.76, d.f. = 1, P =0.0001; Fig. 3a). EMF colonization was slightly higher in roots collected directly from the willow transplants compared to roots from the soil cores (T =1.91, d.f. = 37, P =0.03), which supports the hypothesis that willows were the primary EMF host in this experiment.
Colonization by arbuscules was highly correlated with total AMF colonization (R2=0.82, P =0.0001; data not shown); as trends were statistically similar, we focus on the results for total colonization. AMF colonization of herbaceous roots varied among treatments; however, this effect depended on year (rm-anova: Ftrt = 2.90, d.f. = 4, P =0.02; Table S2). In the first year, AMF colonization of herbaceous roots was significantly lower in WC plots compared to the other four treatments (anova: Ftrt = 108.65, d.f. = 4, P =0.0001; Fig. 3b). In the second year, AMF colonization of herbaceous roots was significantly reduced in MW plots compared to NW plots (planned contrast: F =5.82, d.f. = 1, P =0.02; Fig. 3b). At that time, AMF colonization was comparable in MW and WC plots (planned contrast: F =2.40, d.f. = 1, P =0.12; Fig. 3b). AMF colonization of herbaceous roots was negatively correlated with EMF colonization of woody roots regardless of treatment (Year 1: R2=0.17, P =0.0001; Year 2: R2=0.27, P =0.0001; Fig. 4). None of the environmental principle components were significantly correlated with either EMF or AMF colonization (P >0.1; data not shown).
Effects of Leaf Litter Deposition
The presence of leaf litter increased EMF colonization of woody roots in soil cores collected from the open meadow, but not the willow understorey (litter by habitat interaction, anova: F =12.26, d.f. = 1, P =0.0006; Fig. 5a; see Table S3 in Supporting Information). The positive effect of leaf litter on EMF colonization was consistent across sites and years (four-way interaction, anova: F =0.99, d.f. = 2, P =0.38). In contrast, leaf litter decreased AMF colonization of herbaceous roots, but only in 2009 (litter by year interaction, anova: F =5.14, d.f. = 1, P =0.02; Fig. 5b; Table S3). The negative effect of leaf litter on AMF colonization was consistent across habitats and sites (three-way interaction, anova: F =1.81, d.f. = 2, P =0.17). AMF colonization of herbaceous roots was negatively correlated with EMF colonization of woody roots regardless of habitat or treatment (R2=0.10, P =0.002; data not shown). Leaf litter biomass was not significantly correlated with either EMF or AMF colonization (R2=0.01, P =0.50 and R2=0.01, P =0.38, respectively).
In this study, we used an experimental approach to identify factors driving variation in AMF colonization of herbaceous plants in the field. Above- and below-ground interactions and leaf litter deposition were hypothesized to impact AMF colonization by altering the relative costs and benefits of these associations. Our results indicate that below-ground interactions and leaf litter deposition influence the distribution of mycorrhizal fungi in the krummholz transition zone, which could impact ecotonal boundaries by altering the suitability of these habitats for AM and EM hosts.
Competition for light (above-ground interactions hypothesis) appears to have little effect on AMF colonization, which is not particularly surprising given that alpine plants are often not carbon limited and invest relatively large portions of their carbon budget in root growth (Körner 2003). The shade treatment also did not reduce light availability as much as naturally occurring willows, so this treatment may have underestimated the effect of competition for light on AMF colonization. Other studies indicate that shaded plants have relatively little control over carbon allocation to AMF even when carbon limits plant growth (Olsson, Rahm & Aliasgharzad 2010). Ultimately, reduced irradiance may be more important when other factors also limit photosynthesis, in fast-growing annual plants or in communities where plant density is high (Casper & Jackson 1997).
Our results support the hypothesis that below-ground interactions drive spatial variation in AMF colonization across the willow-meadow ecotone. The absence of a direct willow effect in the NW treatment suggests that the observed below-ground effect was mediated through the willows’ EMF partners; however, the mechanism behind this result remains unclear. Theoretical work argues that resource competition may indirectly affect mycorrhizal associations by limiting photosynthesis and thus carbon supply within the host (Johnson 2010). Field and in vitro studies of interactions between fungal guilds also support the hypothesis that below-ground competition can impact fungal growth (Lodge & Wentworth 1990; Baar & Stanton 2000; Haskins & Gehring 2004; McHugh & Gehring 2006). Finally, several studies acknowledge the potential impact of secondary compounds on mycorrhizal associations and the difficulty of distinguishing between allelopathy and competition (Kovacic, John & Dyer 1984; McHugh & Gehring 2006). Both competition and allelopathy likely play a role in below-ground fungal interactions. Teasing apart the effects of these mechanisms on AMF colonization could shed light on the nature of interaction traits under selection in mycorrhizal associations.
Leaf litter may exacerbate negative below-ground interactions by promoting the growth of microbial competitors. Compared to AMF, EMF are thought to better degrade organic substrates, which could serve as a source of nitrogen (Read & Perez-Moreno 2003). Research also suggests that leaf litter chemistry influences nutrient cycling in a way that maximizes resource uptake by the host’s own mycorrhizal fungi while inhibiting uptake by competing mycorrhizal taxa (Wurzburger & Hendrick 2009). In addition to their own ability to degrade organic substrates, EMF have been shown to translocate resources from saprotrophic fungi to their host plant (Lindahl et al. 1999). All of these mechanisms could give EMF, a competitive edge over co-occurring AMF. Our results support the leaf litter hypothesis and suggest that leaf litter deposition may generate positive plant–fungal feedbacks by promoting the growth of EMF associated with alpine willows. Further work is needed to clarify whether the positive effect of leaf litter on EMF colonization reflects a unique specialization of these fungi for resources in host-derived leaf litter or rather reflects close associations with saprotrophic fungi. Interestingly, leaf litter removal increased AMF colonization in the willow habitat where EMF colonization remained relatively constant. This finding is consistent with the view that leaf litter may influence AMF colonization independent of its effect on EMF and provides tentative support for the hypothesis that phenolic compounds found in willow-derived leaf litter may negatively affect AMF or their hosts (Dudley 2006; Piotrowski, Morford & Rillig 2008). Leaf litter could reinforce the positive feedback between willows and EMF by decreasing the performance of understorey plant or fungal competitors. The role of leaf litter in plant–fungal feedbacks and below-ground interactions may be particularly important in systems like the alpine where much of the accumulated litter is produced by a single dominant host species.
In transplant experiments, unintended disturbance effects or variation in plant community composition can affect response measures. In this case, transplanting willows into the meadow matrix disturbed the soil, which may have altered nutrient availability or disrupted existing fungal networks (Jasper, Abbott & Robson 1991). However, it is unlikely that the observed decrease in AMF colonization in the MW treatment was due to disturbance alone. First, soil nutrient content did not differ significantly among treatments indicating that the transplanting process had little effect on nutrient availability. Second, even though the MW and NW treatments experienced similar levels of disturbance, AMF colonization was only reduced in MW plots. Third, the direction and magnitude of the difference in AMF colonization between NW and MW treatments were similar to the difference between unmanipulated OC and WC treatments and this result mirrors natural variation in AMF colonization across the willow-meadow ecotone (Becklin & Galen 2009). As mycorrhizal colonization depends on the presence of appropriate hosts and plant species richness is generally greater in the open meadow than the willow understorey (Becklin 2010), differences in plant community composition may have also contributed to the observed variation in AMF colonization. However, within each replicate, the MW and NW transplants were c. 3 m apart in previously undisturbed open meadow habitat where the composition of the herbaceous community was similar. Furthermore, we transplanted a standardized pool of potential AM host plants into each plot. If differences in community composition were more important than the presence or absence of EMF, then AMF colonization should have been similar in the two willow transplant treatments. Altogether, the above points argue that neither disturbance nor community composition can fully explain the observed differences in AMF colonization; instead, our results indicate that willows indirectly affect AMF colonization via their EMF partners and leaf litter deposition.
Willow-mediated changes in AMF colonization could have important implications for the fitness and distribution of tundra plants, especially given the upward migration of willows and other treeline species into alpine tundra habitats (Kullman 2002; Elliott & Barker 2004). In this study, we measured colonization at the plot level, which provides insight into how the AMF community as a whole responds to environmental conditions. However, this approach does not account for species-specific responses that could generate variation in plant or fungal diversity and composition across the willow-meadow ecotone. At our field site, AMF composition is similar in open meadow and willow understorey habitats, suggesting that fungal taxa respond similarly to environmental variation in this system (Becklin 2010). Whether these fungi also function similarly across the willow-meadow ecotone and among host species is less clear. Tracing the effects of willow-mediated variation in AMF abundance on the colonization, growth and fitness of individual plants may provide a better understanding of host-specific responses. Unfortunately, measuring mycorrhizal effects on individual alpine plants is challenging due to the slow growth rate and long life span of these perennial hosts.
As shown in this study, below-ground interactions are likely important in habitats where plant density is low, soil resources limit plant growth and neighbouring plants associate with different fungal guilds (Casper & Jackson 1997; Haskins & Gehring 2004; McHugh & Gehring 2006). Alpine ecosystems provide an excellent opportunity to study these interactions at treeline where subalpine EM hosts are encroaching upon alpine tundra communities. In this system, willows are hypothesized to affect plant fitness and species distributions by altering resource availability, alleviating abiotic stress and promoting soil stabilization (Totland & Esaete 2002; Dona & Galen 2006, 2007). Our results indicate that willows may also affect species distributions by altering the suitability of habitats for EM and AM hosts. Positive feedbacks linking willows, EMF and leaf litter deposition could potentially increase the upward distribution of other subalpine EM hosts, while simultaneously restricting the downward distribution of tundra species that associate with AMF. Thus, below-ground interactions could influence ecotonal boundaries and succession in ecosystems, such as the alpine, where EM and AM communities collide.
NSF (DDIG DEB-0808000 and UMEB DBI-0603049; C. Galen), the University of Missouri (K. Becklin), Trans World Airlines (K. Becklin) and the Colorado Native Plant Society (K. Becklin) funded this research. We thank L. Dudley who collected the willow cuttings used for the transplant experiment; T. Cornell, E. Hilpman, J. Han, B. Uelk, P. Marting, J. Geib and A. Michaels helped transplant the willows; R. Callaway, J. Coleman, L. Eggert, C. Gehring, J. Mihail and two anonymous reviewers provided comments on this manuscript.