Foraging Roots and their Associated Microbes Slow the Decay of Belowground Litters in a Temperate Hardwood Forest


 1. There is increasing evidence that plant roots and mycorrhizal fungi, whether living or dead, play a central role in soil carbon (C) cycling. Root-mycorrhizal-microbial interactions can both suppress and enhance litter decay, with the net result dependent upon belowground nutrient acquisition strategies and soil nutrient availability. 2. We measured the net effect of living roots and mycorrhizal fungi on the decay of dead roots and fungal hyphae in a hardwood forest dominated by either sugar maple (Acer saccharum) or white oak (Quercus alba) trees. Root and fungal litter were allowed to decompose within root-ingrowth bags and root-exclusion cores. In conjunction with root effects on decay, we assessed foraging responses and root induced changes in soil moisture, nitrogen (N) availability and enzyme activity. 3. After one year, maple root production increased, and mycorrhizal fungal colonization decreased in the presence of decaying litter Additionally, we found that actively foraging roots suppressed the decay of root litter (-14%) more than fungal litter (-3%), and suppression of root decay was stronger for oak (-20%) than maple roots (-8%). Suppressive effects of oak roots on decay were greatest when roots also reduced soil N availability, which corresponded with reductions in β-glucosidase, (BG) enzyme activity and enhanced oxidative (OX) enzyme activities. 4. These findings further our understanding of context-dependent drivers of root–mycorrhizal–microbial interactions and demonstrate the potential for suppression of soil saprotrophs to contribute to the accumulation of root and fungal derived soil C in temperate forests.


Introduction
Plant roots and their mycorrhizal fungal symbionts actively modify decomposition processes, with important consequences for soil carbon (C) balance. Roots and mycorrhizal fungi can suppress decay and slow C loss by competing with free-living microbes for water and nutrients (Koide and Wu, 2003; Fernandez and Kennedy 2016). The suppression of decay in the presence of mycorrhizal roots or enhanced decay following removal of mycorrhizal roots is termed the 'Gadgil effect' Gadgil, 1971, 1975). Conversely, roots and mycorrhizal fungi can enhance soil C decomposition and loss by releasing exudates to fuel microbial metabolism, known as a 'priming effect' or a 'rhizosphere effect' when microbial enzyme activity is stimulated in root adjacent soils (Kuzyakov 2010). While most studies focus on the occurrence of either a Gadgil or priming effect, these context-dependent processes are not mutually exclusive (Fernandez and Kennedy 2016). To infer the net effect of mycorrhizal roots on decay processes we need a better understanding of which plant, soil, and litter characteristics promote Gadgil over priming effects and vice versa. Currently this knowledge gap hinders progress in incorporating rootmicrobe interactions into ecosystem models (Moore et al. 2015), despite increased recognition that roots, and mycorrhizal fungi are important drivers of soil C dynamics (Jackson et al. 2017).
Tree species differ in root functional traits related to belowground nutrient acquisition strategies, which in turn, may have differential effects on decomposers (Han et al. 2020). Belowground foraging strategies can range from 'do-it-yourself' acquisition to 'outsourcing' acquisition to mycorrhizal foraging (Bergmann et al. 2020). Plants that use a DIY strategy prioritize e cient root construction by making thin roots with high speci c root length (SRL). This may lead to greater competition for resources between roots and microbes. Under these conditions, reduced rates of nutrient mineralization caused by fewer or less active soil microbes might be offset by ne root systems that can proliferate rapidly into nutrient patches (Hodge 2004;Liese et al. 2017). In contrast to a DIY strategy, roots might outsource resource acquisition to mycorrhizal symbionts, which result in greater resource competition between mycorrhizal fungi and soil saprotrophs. This leads to the prediction that heavily colonized roots or root systems with an increased number of mycorrhizal tips may be associated with (or lead to) stronger Gadgil effects. As an alternative outsourcing strategy, plants may invest C in root-associated microbial communities and enzymatic activity rather than selectively investing in mycorrhizal fungi. Thin absorptive roots with high SRL have been shown to enhance rhizosphere effects ; Han et al. 2020). Currently it is unclear the degree to which root traits and foraging strategies predict root-mycorrhizal-microbial interactions.
Mycorrhizal fungi are an extension of a plant's nutrient acquisition strategy, and the type of mycorrhizal fungi a plant associates with can in uence the direction and magnitude of root effects on decay . The Gadgil effect is typically associated with ectomycorrhizal (EcM) fungi, as some species of EcM fungi produce oxidative enzymes to mobilize nitrogen (N) from organic matter (OM), deterring further decay by free-living saprotrophs (Shah et al. 2016;Zak et al. 2019). In contrast, arbuscular mycorrhizal (AM) fungi do not produce the oxidative enzymes required to degrade OM, though they may modify the activity of free-living saprotrophs via water and nutrient reduction or through hyphal turnover and exudation (Bunn et al. 2019 Meier et al. 2017). Whether mycorrhizal roots inhibit or facilitate other microbes in low nutrient soils likely depends on the nutrient content or quality of litter inputs themselves (Smith and Wan 2020). It is thought that higher quality litter inputs (lower C:N ratios), in the earlier stages of decay increase carbon use e ciency among free-living decomposers and can limit the advantage of EcM fungi in acquiring organic N, weakening the Gadgil effect (Fernandez et al. 2020). More studies are needed to disentangle the effects of soil and litter nutrient conditions on mycorrhizal interactions with saprotrophs.
We sought to evaluate the net effect of living roots and mycorrhizal fungi on decaying roots and fungi differing in their resource quality within forest stands differing in mycorrhizal type and soil nutrient conditions. We pose the following questions (i) Do root traits related to foraging (i.e., root production, morphology, and mycorrhizal colonization) re ect root involvement in decomposition? (ii) How do living roots and mycorrhizal fungi in uence the decay of dead roots and fungi in a hardwood forest for both an EcM (white oak) and AM-associated (sugar maple) tree species? We paired root ingrowth bags with root exclusions (See Fig. 1) to measure root foraging responses and effects on decomposition (mass loss). In addition to root and fungal mass loss, we measured how roots altered the availability of limiting resources (e.g., water and nitrogen), N cycling rates and rhizosphere enzyme activity to test for the occurrence of a 'priming' effect.
We hypothesize that (i) trees respond to litter additions by altering belowground foraging behavior (by either increasing root proliferation or mycorrhizal colonization) and (ii) the presence of mycorrhizal roots in uences mass loss from root and fungal litter. Further, we hypothesize that the magnitude of root effects on decay differs between tree species and litter types. We predict that oak roots growing in Nlimited soil environments will be associated with a stronger Gadgil' effect than maple roots growing in soils with increased inorganic N and faster rates of N cycling. We predict that the Gadgil effect will be reduced for higher quality fungal litter compared with lower quality root litter (Fig. 1).

Site Description
We conducted this work at Moores Creek, a mature hardwood forest located in south-central Indiana EcM-associated trees species comprised greater than 85% of stand basal area. AM plots include a mixture of the following species: Acer saccharum, Lirodendron tulipifera, Prunus serotina and Sassafras albidum. EcM plots contain a mixture of Carya glabra, Fagus grandifolia, Quercus alba and Quercus rubra. We chose three plots in which Acer saccharum (hereafter referred to as "maple") was the predominant AM-associated canopy species and another three plots in which Quercus alba (hereafter referred to as "oak") was the predominant EcM-associated canopy species. We chose these two species because they are dominant species across much of the Northeast and Midwest Within each plot, we selected four trees (12 trees total per species). Paired oak and maple plots were organized in three forest blocks according to location (Fig. S1).

Decomposition Measurements
We collected species-speci c root litter in September of 2017, for additional details see SI methods. To generate fungal hyphae, we grew Melinomyces bicolor hyphal plugs in potato dextrose broth for approximately 30 days. After 30 days, we rinsed the M. bicolor hyphae with DI water and then dried it 20 o C for 48 hours (Fernandez and Kennedy 2018). We chose M. bicolor because it is a melanized ectomycorrhizal fungal species which has been shown to have mass remaining after three months of decay (Fernandez and Kennedy 2018). Dominant EcM fungal species at the site include Cortinarius caperatus, Elaphomyces decipiens, Hygrophorus sordidus and Russula ochroleuca (Beidler et al. 2020).
To determine initial litter C: N values, we ground dried root and fungal material to a powder using a GenoGrinder (SPEX® SamplePrep) and analyzed the ground material for total C and N (Elemental Combustion System 4010; Costech Analytical Technologies, Valencia, CA, USA). We constructed litter bags (4.5 cm x 4.5 cm) from 0.2 mm nylon mesh (4.5 cm x 4.5 cm) and lled them with either 100 mg of root litter for each tree species or 25 mg of M. bicolor fungal hyphae. We determined that these masses were large enough to detect signi cant changes in mass loss (Träger et al. 2017;Beidler et al. 2020) and could serve as realistic proxies for organic nutrient patches within the soil environment (Hodge 2004).
We incubated root and fungal litters in two environments -inside root ingrowth bags and root exclusions (details below; Fig. 1). We constructed root ingrowth bags (20 cm x 25 cm) from polyester fabric with a 0.5 mm mesh size (Eissenstat et al. 2015). During the rst week of May 2018, we tracked a live woody root (> 3mm in diameter, ~ 25 cm in length) to a focal tree (at a depth of 10-15 cm) and carefully removed all lateral absorptive roots. We pruned roots to initiate new absorptive root growth and to measure root foraging responses to litter decay. For each tree, we placed root and fungal litter in half of the root bags, on either side of a pruned root (Fig. 1). We then lled root bags with root free soil or soil picked free of roots and passed through a 2mm sieve (~ 300g) collected from around each tree (10-15 cm depth). We collected a subsample of soil to determine initial soil pH and total soil C and N for each focal tree (Fig. S1). We installed two sets of root bags (one bag with litter and one bag without litter) to allow for two separate harvests (3 months after installation and 1 year after installation). In total we installed 48 root bags per tree species (4 root bags x 4 trees x 3 plots).
We made root exclusions by driving PVC pipe (15.2 cm diameter) to a 30 cm depth. We then removed the PVC pipe (keeping the soil column inside intact) and wrapped the bottom with 1 µm mesh to exclude roots and mycorrhizal hyphae. We left exclusion cores to equilibrate for a year (n = 6, 1 per plot). At the same time, we buried litter bags at a similar depth inside of exclusions. In August of 2018 (after 3 months) and May of 2019 (after 1 year), we collected litter bags and dried them at 60°C for 48 hours to determine mass loss. At the time of harvest, we removed any living roots colonizing litter bags.
To test for the potential disturbance effects of soil sieving and an additional layer of mesh fabric on mass loss from litterbags within root bags vs. exclusions, we conducted a follow up decomposition experiment where we decayed litter inside (enclosed litter bag) and outside of root bags (non-enclosed litter bag) within PVC exclusions. We found no differences in mass loss for litter decaying inside of the enclosed and non-enclosed litter bags (p = 0.560, Table S1), indicating that bag environment did not signi cantly affect mass loss.

Root trait measurements
We harvested root bags by block in the same order that we installed them. During processing we excavated and carefully cleaned the intact root networks within root bags. We excluded any root bags containing dead roots from subsequent analyses (see Table S2 for information on pruning recovery).
Following cleaning, we removed absorptive roots (orders 1-2) from transportive roots (3rd order and higher) using stream-order classi cation (Guo et al. 2008). We oated roots in a transparent tray and imaged them using an Epson Expression 100000XL scanner (300 dpi). We analyzed images for root length, average diameter, root volume, and tip number using Win-RHIZO (Regent Instruments Inc 2009).
We then dried and weighed roots (60ºC for 48 hours) to determine root tissue density (g cm − 3 ) and speci c root length (m g − 1 ).

Soil measurements
To minimize disturbance of soil inside of exclusions, we took point measurements of soil volumetric water content (VWC) using a Hydrosense II soil-water sensor (Campbell Scienti c). At the end of the study (365 days), when root bags were fully colonized by living roots and pruning related disturbances had likely subsided, we sampled soils from root bags and exclusions for differences in nitrogen (N) availability and enzyme activities using the methods of Brzostek et al. (2015) and Midgley and Phillips (2019). We extracted soil inorganic N (NH4+-N and NO3-N; µg g − 1 ) from a 5g soil subsample using 10mL of a 2M KCl solution. We determined N mineralization and nitri cation rates (µg g − 1 d − 1 ) using an additional set of soil subsamples which we incubated for two weeks at 25° C, subsequently extracted with 10 mL of 2 M KCl solution and analyzed on a Lachat QuikChem 800 Flow Injection Analyzer (Lachat Instruments, Loveland, CO, USA).
We performed assays for the following enzymes: b-1,4-N-acetylglucosaminidase (NAGase-involved in N degradation), β-Glucosidase (BG; involved in degradation of labile C), oxidative (OX) enzyme activity (the sum of peroxidase and phenol oxidase activities; enzymes involved in the degradation of complex C). We measured potential enzyme activities for soils within root bags and PVC exclusion cores, as well as for rhizosphere soils in root bags with and without litter additions. We performed both colorimetric (OX) and uorometric (BG and NAG) assays by preparing soil slurries (1.5 g soil and 100 mL of sodium acetate buffer per sample: pH = 5) and adding them to microplates in triplicate. To determine BG and NAG activities, we added methylumbelliferone (MUB) substrate, incubated the microplates in the dark at 23°C for either two (NAG) or ve (BG) hours and read the plates using a microplate uorometer (365 nm excitation and 450 nm emission). To determine OX enzyme activities, we added L-3,4dihydroxyphenylalanine (L-DOPA) substrate, incubated microplates in the dark at 23°C for four hours and read the plates using a microplate spectrophotometer (absorbance at 460 nm). We corrected potential enzyme activities (µmol g − 1 h − 1 ) for controls, quenching, and dry soil weights (Midgley and Phillips 2019).

Calculations and statistical tests
We calculated root mass loss by subtracting ash-free dry mass of the remaining litter from the ash-free dry mass of the initial input for each litter bag. We calculated pruning recovery as the percentage of woody roots that grew new absorptive roots following pruning (Eissenstat et al. 2015). We calculated speci c root length 1 + 2 (SRL), as the length of the 1st + 2nd order roots divided by the corresponding 1st + 2nd order root dry weight. We calculated root tissue density 1 + 2 (RTD) as 1st + 2nd order root dry weight divided by1st -2nd order root volume. We calculated root branching intensity as the number of tips per 1 + 2nd root length (Liese et al. 2017). We determined total root length by combining measurements for the absorptive and transportive roots and dividing by the duration of the growth period to determine new root length production (cm d − 1 ) following pruning. We determined the percentage mycorrhizal colonization for oak roots by dividing the number of rst-order roots colonized by EcM fungi by the total number of rst order roots examined; to determine AM colonization intensity for maple roots we used the grid line intersect method (Brundrett et al. 1996). Five maple root samples were damaged during clearing and omitted from mycorrhizal colonization calculations.
We determined the magnitude of root effects for each plot by subtracting measurements taken from exclusions from root bag treatments to calculate root induced changes in mass loss, soil moisture content, N cycling and enzyme actives (Fig. 1). We determined the magnitude of litter effects on root foraging for a given tree by subtracting measurements taken from root bags without litter additions from root bags with litter additions to calculate the litter induced changes in root foraging traits (Fig. 1). We performed one sample T tests to determine if mean changes differed from zero.
Because mass remaining, and VWC data are bounded by zero and one, they often did not meet the assumptions of normality and/or homoscedasticity. As an alternative to transformation, we t generalized linear mixed-effect (GLME) models based on a beta distribution with a logit link function using the glmmTMB package in R (Douma and Weeden, 2019). Fixed effects included root treatment, tree species, incubation time and forest block. Plot nested within block was included as a random effect. We included soil VWC to control for potential moisture differences between treatments and plots. We also included treatment, species, and time interactions. We performed type III Wald chi-square tests to determine statistical signi cance of GLME models (α = 0.05; Anova.glmmTMB function).
We assessed the effect of root treatment, tree species and potential interactions on soil variables using linear mixed-effect (LME) models with plot nested in block as a random factor. We also used LME models to test the xed effects of litter addition, tree species, growth period and all interactions on root foraging traits. We included tree nested in plot as a random effect. We used the ANOVA function in the car package to generate analysis of deviance tables for LME models (Type III Wald chi-square tests; α = 0.05; Fox, 2019; See SI for ANOVA tables). We explored interactions using Tukey post hoc tests in the lsmeans package (Lenth, 2018). We performed regression analysis to test for signi cant relationships between correlated continuous variables (α = 0.05). We carried out all statistics using R 4.0.3 (R Core Team, 2020).

Root Responses to Presence of Decaying litter
We found that morphological root traits did not differ between tree species, except for average diameter 1 + 2 and tissue density 1 + 2 (Table 1). First and second order oak roots were thinner and denser on average. New root production was higher for oak roots during the rst three months (oak, 2.5 ± 0.3, maple, 1.3 ± 0.2 cm d − 1 ; t = -1.6, p < 0.001) but did not differ between species after a year (t = -0.11, p = 0.9; Table S3). Overall, recovery from pruning or the percentage of roots that grew new absorptive root length, was greater for roots grown in soil with litter additions (79%) when compared to soils without litter additions (60 %; χ 2 = 4.3, p = 0.04; Table S2). On average root trait values did not differ between soil treatments for either sampling date (Table 1).
Signi cant litter induced changes in root production and mycorrhizal colonization were detected for maple roots after 365 days of growth (Fig. 2). After one year, maple root production was 70% higher and mycorrhizal colonization was 8% lower in the presence of decaying litter (Table S4). After one year of root ingrowth, rhizosphere enzyme activities were similar between tree species and soil treatments, except for NAGase, which was twice as high in oak rhizosphere soils in the presence of decaying litter (Table 2). Moreover, for roots growing in soils with litter additions, rhizosphere activity of C-degrading enzymes related negatively to root mycorrhizal colonization (slope: -0.013; R 2 = 0.23; p = 0.03) and positively to root production (slope: 0.37; R 2 = 0.21; p = 0.03). Across species, heavily colonized mycorrhizal roots tended to have lower oxidative enzyme activities in rhizosphere soils, while rhizosphere β-Glucosidase (BG) activity increased with root production (Fig. 3).

Root Effects on Decay
We found that the presence of foraging roots suppressed decay and that the magnitude of root reductions in litter mass loss depended on tree species, time period, and litter type (Fig. 4a). The suppression of decay was greater for oak root litter. When averaged across time periods, mass loss of oak root litter decreased by 20% and maple root litter decreased by 8% in the presence of living roots. For both oak and maple trees, root effects on decay decreased over time (Fig. 4b), supported by a signi cant treatment by time interaction with respect to root mass remaining (χ 2 = 7.2, p = 0.007; Table S8). When averaged across species, root mass remaining in exclusions after three months (62.1 ± 2.9 %) was similar to root mass remaining in the presence of living roots after one year (67.9 ± 1.8 %; t = -1.9; p = 0.2).
Relative to root litter, fungal litter decomposed rapidly, with 85% of mass loss occurring in the rst 92 days, more than triple that of root litter (~ 25% mass loss after 92 days; Fig. 4b). However, the only signi cant root induced change in fungal decay was a 6% decrease in fungal mass after 365 days in the presence of maple roots (t = -3.02; p = 0.02; Table S8) Exclusion of roots and mycorrhizae also altered soil conditions, with the direction of root induced changes differing between tree species. Across time periods the presence of oak roots increased soil VWC (+ 2.1 ± 0.4 %; t = 4.78, p < 0.001) while the presence of maple roots decreased soil VWC (-4.0 ± 1.1%, t =-3.4, p = 0.004; Table S8). After 1 year, the presence of roots altered soil N availability without altering soil pH (Table 3). Oak roots decreased soil inorganic N (-4.7 ± 6.9 µg g − 1 ; t = -2.26, p = 0.05 while maple roots increased soil inorganic N (+ 9.12 ± 4.5 µg g − 1 ; t = 5.39, p = 0.002; Table S9). Differences in soil inorganic N concentrations were likely due to differences in N cycling rates in the presence of roots (Table 3). Nitri cation rates were 5.5 x lower and N mineralization rates were 2.5 x higher when oak roots were present. Whereas, nitri cation rates were 4 x higher and N mineralization rates were 7 x higher when maple roots were present. However, it is important to note that when roots were excluded, average soil inorganic N concentrations were similar between maple and oak plots (t = 0.89, p = 0.80; Table 3).
We also detected signi cant oak root induced changes in carbon (C) degrading enzyme activities (Fig. 5).

Discussion
Approximately 50-70% of soil C is thought to originate from belowground inputs (Godbold et al. 2006;Clemmensen et al. 2013), yet the degree to which roots and mycorrhizal fungi in uence soil C storage is poorly understood in temperate forests. We hypothesized that trees would respond to root and fungal litter additions by altering root morphology or increasing mycorrhizal colonization, and that the magnitude of root effects on decay would differ between tree species and litter types. We found that maple roots grew faster and were less colonized by AM fungi in the presence of decaying litter (Fig. 2). We also found that across species, increased root production in the presence of decaying litter was associated with higher rhizosphere BG activity and heavily colonized mycorrhizal roots were associated with lower rhizosphere OX activities (Fig. 3). However, root foraging traits and rhizosphere enzyme activities did not relate directly to root induced changes in decay. We predicted that suppressive effects of mycorrhizal roots on decay would be greater for low quality litters (roots) than high quality litters (fungi), and greater for oak roots (which associate ECM fungi) than maple roots (which associate with AM fungi). We found that living roots and mycorrhizal fungi inhibited the decay of root litter to a greater degree than fungal litter (Fig. 4.), and the magnitude of this effect was greater for oak roots growing in N limited soils which also tended to have reduced BG activity and enhanced activity of OX enzymes (Fig. 6). Collectivity, ) our results show that living roots can play a leading role in the carbon and nutrient dynamics of decaying roots and the potential for litter quality and tree species identity to modify that role.

Root Responses to Presence of Decaying litter
Maple and oak roots differed in their foraging response to the presence of litter. Maple root production increased, and mycorrhizal colonization decreased (Fig. 2), consistent with other root pruning studies (Eissenstat et al. 2015). Maple roots are believed to forage more by proliferating roots rather than by allocating resources to mycorrhizal fungi, re ecting a 'DIY' strategy (Bergmann et al. 2020). If greater root proliferation decreased nutrient or water availability for free-living microbes, this might explain the slight suppression of litter decay by maple roots. Faster growing roots were associated with increased rhizosphere BG activity (Fig. 3), indicating that actively proliferating roots might also stimulate, rather than suppress microbial degradation of C in the soils immediately surrounding absorptive roots. This nding provides some support for a new paradigm of plant-nutrient acquisition proposed by Sun et al. (2021), in which increased root exudation might complement morphologically competitive root traits, allowing plants to promote soil nutrient availability instead of waiting for nutrients to become available (Yin et al. 2018). More studies are needed to understand the spectrum of belowground competitive strategies and how they might relate to decay processes.
Oak root foraging traits did not respond to litter additions (Table 1) and this lack of root response might be indicative of greater reliance on mycorrhizal foraging away from roots (Chen et al. 2018;Cheng et al. 2018). It is thought that EcM fungal species with long distance exploration types or hyphae that span larger distances from root tips have greater enzymatic capabilities and thus might contribute to a stronger Gadgil effect in hyphosphere (soils surrounding hyphae) rather than rhizosphere soils (Tedersoo and Smith 2013). Furthermore, root mycorrhization was negatively related to oxidative enzyme activity in rhizosphere soils, indicating that mycorrhizal fungi were likely not enhancing OX activities in root adjacent soils (Fig. 3). Instead, saprotrophic fungi (Baldrian 2008) or even bacteria (Lladó et al. 2017) might have been the primary producers of oxidative enzymes in rhizosphere soils. Future studies should relate enzyme production to fungal community composition at increasing distances from the root to better understand competitive dynamics between mycorrhizal fungi and free-living soil saprotrophs.

Root Effects on Decay
Our results suggest that foraging mycorrhizal roots actively alter soil water availability, N cycling and enzyme activity. Ingrowth of maple roots reduced soil moisture (Fig. 4a) which might have contributed to the suppression of maple root decay (Koide and Wu, 2003). Surprisingly, the presence of oak roots increased the moisture content of surrounding soils. Hydraulic lift or redistribution of water from deeper soil layers to oak rhizosphere soils during summer months might explain this result (Ishikawa and Bledsoe, 2000). Moreover, white oak (Q. alba) trees are deeper rooted and might depend more on deep water sources compared to sugar maple (A. saccharum) trees (Lanning et al. 2020). Oak and maple roots both increased N mineralization rates; however, maple roots increased nitri cation rates whereas oak roots decreased nitri cation rates (Table 3). These results are consistent with previous studies reporting elevated nitri cation in maple-dominated stands compared to adjacent oak-dominated stands (Vitousek et al. 1982;Lovett and Mitchell, 2004). The reasons for the differences are unclear but may relate rootpromotion of distinct microbial communities. In a study conducted on the same site, Mushinski et al, (2021) reported that plots dominated by AM trees (many of which were maple-dominated) had soil microbial communities with 4-fold more N cycling genes than ECM-dominated plots (many of which were oak dominated). It also suggests that maple roots may have been less limited by N and instead could have been competing for water or phosphorus (DeForest and Snell 2019).
Given that oak roots stimulated N cycling to a lesser degree than maple roots, N availability may have contributed to a stronger Gadgil effect. Root-induced changes in N cycling in oak plots corresponded with reductions in the activity of β-glucosidase (Fig. 6), suggesting that the cost of synthesizing N-rich enzymes may have slowed saprotrophic activity. Mycorrhizal hyphae, especially those produced by EcM fungi, extend beyond the rhizosphere into the surrounding soil, and mycorrhizal hyphae independent of roots have been shown to reduce BG activity and slow root litter decay (Lin et al. 2019). At the same time, EcM fungi can accelerate N cycling via priming effects (Meier et al. 2015). Thus, the accelerated rates of N mineralization in the oak soils (Table 3)  The labile nature of fungal hyphae (6 X more N per unit C) when compared with the root litter may explain the differential effects of living roots on root and fungal decay in this study. We detected a suppressive effect of maple roots on fungal mass loss after a year and though not statistically signi cant, the effect of oak roots showed a similar trend. The fungal hyphae used in this study contained signi cant concentrations of the chemically complex pigment, melanin (Fernandez and Koide 2014). It is possible that at this stage in decay, the fraction of fungal mass remaining was resistant to decay and there was a shift in the decomposer community to less e cient decomposers (Fernandez and Kennedy 2018). A previous study conducted at this site found that in plots dominated by maple trees but containing EcMassociated understory vegetation, the proportion of EcM fungi colonizing decaying M. bicolor fungal hyphae increased through time (between 14 and 92 days; Beidler et al. 2020). To account for decomposer community shifts associated with substrate quality, we suggest that future studies testing 'Gadgil effects' on fungal decay include increased sampling frequency to capture both shorter-and longer-term time intervals.

Limitations and Conclusions
We found that root production and mycorrhizal colonization were responsive to belowground litter decay and related to rhizosphere enzyme activity. However, we only measured the foraging behavior of two tree species. To develop a generalizable understanding of belowground nutrient acquisition strategies, more studies are needed to test for trait coordination between roots and root-associated microbes for a greater diversity of plant species (Chen et al. 2018;Sun et al. 2021). We show that the Gadgil effect is not speci c to EcM forming gymnosperms in boreal forests or leaf litter decaying in organic soil horizons. We found that both oak and maple tree roots can have a suppressive effect on decay and that the magnitude of this effect was greater for low-quality root litter decaying in N limited mineral soils. One caveat of our design is that we pruned roots to ensure that they were actively growing and placed them in root-free, sieved soils which may have enhanced root induced changes in decay compared to undisturbed soil, especially early on in decay. However, our ndings demonstrate the potential for the Gadgil effect to operate on root and fungal inputs in temperate forests and to contribute to the persistence of root and fungal derived C in forest soils. Given the known sensitivity of root-microbe interactions to global change  Experimental design and litter carbon: nitrogen values. Litter induced changes in root foraging behavior (morphology and growth) were assessed after 92 and 365 days of root ingrowth. Root (and mycorrhizal fungal) induced changes in litter mass loss were determined after 92 and 365 days of decay for Quercus alba (Oak), Acer saccharum (Maple) and Melinomyces bicolor (fungal) litter. For each litter type C and N values are presented as mean values ± SE (n=3). For a particular C or N value, means not sharing a letter are signi cantly different (P < 0.05).

Figure 2
Litter-induced changes (∆LI) in Quercus alba (Oak) and Acer saccharum (Maple) (a) root length production and (b) mycorrhizal colonization rates at 3 months (92 days) and 1 year (365 days). Error bars represent 95% con dence intervals (CIs). One sample T-tests were performed to determine whether root induced changes differed signi cantly from zero (denoted by an asterisk*; see Table S5 for T-test results).  Loss), as well as soil volumetric water content (Soil VWC) at 3 months (92 days) and 1 year (365 days). Error bars represent 95% con dence intervals (CIs). One sample T-tests were performed to determine whether root induced changes differed signi cantly from zero (denoted by an asterisk*: see Table S9 for T-test results) (b) Average mass remaining ± SE of root and fungal litter through time. Root (and mycorrhizal fungal) induced changes in soil organic matter (SOM) degrading extracellular enzyme activities in soils dominated by either oak (Quercus alba) or maple (Acer saccharum) roots on the nal sampling date (365 days). Data is presented for the following enzymes: β-Glucosidase (BG), b-1,4-Nacetylglucosaminidase (NAG), and oxidative (OX) enzymes (phenol oxidase and peroxidase activities summed). Error bars represent 95% con dence intervals (CIs). One sample T-tests were performed to determine whether root induced changes differed signi cantly from zero (denoted by an asterisk*; see Table S12 for T-test results).