Global mycorrhizal fungal range sizes vary within and among mycorrhizal guilds but are not correlated with dispersal traits

Mycorrhizal fungi associate with the majority of plant species with large consequences on ecosystem nutrient, carbon and water cycling. Two main types of mycorrhizal fungi, arbuscular mycorrhizal (AM) fungi and ectomycorrhizal (EM) fungi, dominate terrestrial ecosystems. Most global distribution modelling of AM and EM associations describe the distribution of AM and EM plants, and not fungi directly. However, significant functional trait variation occurs within AM and EM fungal guilds. Therefore, modelling range sizes and determinants of these ranges of fungi directly is likely to create spatial maps that are a better proxy of ecosystem function than guild‐level lumping of AM and EM plant distributions.

mycorrhizal (AM) fungi and ectomycorrhizal (EM) fungi, dominate most terrestrial ecosystems. AM fungi originated ca. 480 million years ago, whereas EM fungal lineages have evolved multiple times from saprotrophic fungal clades ca. 183 million years ago (Lutzoni et al., 2018). These mycorrhizal fungal types also differ in function, with AM fungi largely accessing inorganic nutrients in soils while EM fungi can decompose organic forms of N and P, leading to different nutrient use strategies among plant hosts of the two groups (Averill, Bhatnagar, Dietze, Pearse, & Kivlin, 2019). In addition, while host specificity of mycorrhizal fungal associations are still unresolved (Opik & Peay, 2016), general differences are recognized between AM and EM fungi. Many AM fungi lack specific plant host associations (Veresoglou & Rillig, 2014) and associate with at least 70% of plant taxa (Bueno et al., 2017;Soudzilovskaia et al., 2019). In contrast, EM fungi can specialize at the level of plant genera or even species (Matheny et al., 2009;Tedersoo et al., 2008;Toju, Yamamoto, Tanabe, Hayakawa, & Ishii, 2016) and associate with approximately 2% of plant taxa (Tedersoo & Brundrett, 2017). In addition, species concepts vary among AM and EM fungi. AM fungi are typically grouped into virtual taxonomic units (VTXs) based on 97% sequence identity in the highly conserved SSU gene region sensu Opik et al. (2010), which may represent genetic variation at coarser taxonomic levels than species (Bruns & Taylor, 2016) and artificially inflate "species" ubiquity. EM fungal species delineations are more resolved via grouping at 97% sequence identity of the variable ITS2 locus (Koljalg et al., 2013), which may result in narrowed "species" ranges. Finally, most AM fungi are probably asexual (Kokkoris & Hart, 2019) and therefore their range sizes may not be hindered by lack of suitable fungal mating types. In contrast many EM fungi, especially in the Basidiomycota, require two compatible mating types in order to create viable lineages of sexual spores (Horton, 2017).
There is also substantial functional variation among taxa within AM-and EM fungal guilds. For example, AM fungi in the Gigasporaceae and Glomeraceae generally produce more hyphae outside of roots (in soils) versus inside of roots respectively (Hart & Reader, 2002;Powell et al., 2009). Trade-offs in growth form among AM fungal lineages may result in functional variation in resource competition, disturbance tolerance or stress tolerance (Chagnon, Bradley, Maherali, & Klironomos, 2013). For example, fungi in the Diversisporaceae have greater access to soil nutrients and fungi in the Glomeraceae tend to compete with pathogens for space in the root, resulting in different benefits for the host plant Treseder et al., 2018). Additional trade-offs may occur for resource acquisition of nitrogen versus phosphorus among AM fungal lineages (Treseder et al., 2018). Similar distinctions occur among EM fungal lineages where some clades (e.g. Lactarius spp.) excel at overall decomposition of organic matter (Baldrian, 2006;Martin, Kohler, Murat, Veneault-Fourrey, & Hibbett, 2016) and others are specialized to decomposition of specific organic substrates (e.g. lignin or tannin) (Martin et al., 2016;Shah et al., 2016;Talbot, Martin, Kohler, Henrissat, & Peay, 2015). Moreover, some trait trade-offs follow more traditional expectations from plant-based ecological theory (Grime, 1979), including clades that can be more competitive (e.g. Helvella spp.) and others (e.g.Suillus spp.) that can disperse farther within local ecosystems (Smith, Stedinger, Bruns, & Peay, 2018).
The large functional variation between and within AM and EM fungal guilds has led to many studies of how mycorrhizal fungal guilds might impact ecosystem functions such as carbon storage, nutrient cycling and litter decomposition across space (e.g. Averill One of the main initial processes controlling the distribution of many microorganisms is dispersal limitation (Hanson, Fuhrman, Horner-Devine, & Martiny, 2012). While historical paradigms of microbial biogeography suggested that dispersal limitation did not apply to microorganisms (Baas Becking 1934), more recent evidence suggests that spore traits (Kivlin, Winston, Goulden, & Treseder, 2014), wind patterns (Tipton et al., 2019), seasonality , life history strategies (Horton, 2017) and the location of spore origination (Egan, Li, & Klironomos, 2014;Peay, Garbelotto, & Bruns, 2010a;Peay, Kennedy, Davies, Tan, & Bruns, 2010b;Peay, Schubert, Nguyen, & Bruns, 2012) all affect how far fungi will disperse in the atmosphere. Thus, dispersal limitation should affect range sizes at the global scale assuming airborne passive dispersal is the main dispersal vector of mycorrhizal fungi and any of these factors influence how far a fungal propagule disperses. Alternatively, spore size may be inversely related to dispersal distance (sensu Davison et al., 2018) if larger spores are more likely to germinate (Norros, Karhu, Norden, Vahatalo, & Ovaskainen, 2015) or withstand resource limitation (Halbwachs, Hielmann-Clausen, & Bassler, 2017). If mycorrhizal fungal spores are actively dispersed, then metrics of dispersal limitation (e.g. spore size) that affect passive dispersal processes should not be related to mycorrhizal fungal range sizes.
Here I address how spore size and phylogenetic history independently and jointly affect range sizes of AM and EM fungal taxa. I collected data on geographic distributions of AM and EM fungal taxa at the global scale to determine taxon-specific range sizes and the influence of dispersal limitation on those ranges. If dispersal limitation affects range sizes, I expected EM fungi to have larger ranges than AM fungi, given EM fungal spores are mainly produced by mushrooms aboveground and are smaller (79-20,100 μm 3 ) than AM fungal spores (154,000-55,750,000 μm 3 ). Alternatively, if spore size is not related to range size, I expected AM fungi to have larger ranges given that they originated over 200 million years before EM fungi and have a wider plant host breadth than EM fungi.

| Data collection
AM fungal occurrences were collected from the MaarjAM database (Opik et al., 2010) and supplemented with a GenBank search using

| Range sizes
Commonly applied estimates of range sizes, such as convex hulls, can overpredict range sizes by including gaps in distributions (Burgman & Fox, 2003). To mitigate errors in range size calculations, I calculated range sizes in several ways. First, I calculated range sizes using alpha hulls in the range Builder package v 1.4 (Rabosky et al., 2016) where hulls included 95% of datapoints and were trimmed to include only terrestrial areas followed by poly- Northern hemisphere).

| Phylogeny
Separate phylogenies were created for AM and EM fungi using representative sequences for each VTX and OTU respectively. For each guild, sequences were trimmed to equal length and aligned with MAFFT and placed into a maximum likelihood phylogeny using RAxML (Appendix 1). This process was iterated until the best alignment and phylogeny were obtained using PASTA (Mirarab et al., 2015). Sequences for AM fungi were derived from the 18S gene sequences available in MaarjAM (Opik et al., 2010). Sequences for EM fungi were obtained from Tedersoo et al. (2014) from the ITS2 variable region. While variable, the ITS2 phylogeny was monophyletic with the exception of one Russula species which grouped with a sister genus Lactarius. Because these genera are closely related within the same family, this is unlikely to influence the overall phylogenetic model (see below).

| Spore size traits
Spore sizes (height, width, length) for AM fungi were obtained from Aguilar-Trigueros, Hempel, Powell, Cornwell, and Rillig (2019). When VTX were not represented, I used the mean value at the genus level.
The same EM fungal spore traits were collated from the primary literature (see Appendix S2). Spore volumes were then calculated based on the volume of an ellipsoid.

| Statistics
Differences in range size between each mycorrhizal fungal guild were determined using a univariate ANOVA with the fixed factor of mycorrhizal fungal type. Because hyphal exploration types, nutrient acquisition and response to global change are often conserved within mycorrhizal fungal genera (Chagnon et al., 2013;Phillips et al., 2019;Treseder et al., 2018), I also used univariate ANOVAs to determine how spore size and range sizes varied among genera within each mycorrhizal fungal guild when a genus contained at least three representative fungal taxa.
I tested if spore volume or range size metrics were phylogenetically conserved using Blomberg's K in the Picante package v 1.7 in R (R Core Team, 2018). Phylogenetic signal was tested separately for AM and EM fungi since the phylogeny for each group was created with a different marker gene.
My main goal was to understand if spore volume was related to any metric of range size. I tested for these univariate relationships separately for EM and AM fungi with phylogenetic generalized least squares (PGLS) approach to account for phylogenetic signal in fungal traits using the Caper package v 1.0.1 (Orme et al., 2012) in R (R Core Team, 2018).

| Range sizes
Overall, AM fungi had 3.43x larger ranges by area (F = 192.400, p < .001, Figure 1a) and 1.44× larger latitudinal ranges than EM fungi (F = 130.200, p < .001, Figure 1b). Because species definitions may vary among fungal guilds, I also calculated range size for each EM fungal genus. Even in this case, AM fungal taxa had larger areal range sizes (F = 4.487, p = .035), but not latitudinal ranges (F = 1.848, p = .175), than EM fungal genera.

| Phylogenetic signal in spore and range size traits-genus level
Within guilds, spore sizes varied among AM fungal genera (F = 37.73, p < .001, Figure 3a)

| Relationships between spore size and range size
Despite AM fungi having larger ranges and larger spore sizes than EM fungi, within both guilds spore size was unrelated to range size calculated by area (AM R 2 = .006, p = .961; EM R 2 = .037, p = .082, Figure 4a,b) or by latitudinal range (AM R 2 = .006, p = .768; EM R 2 = .023, p = .135) after phylogenetic correction. However, EM fungal taxa with smaller spores tended to have larger latitudinal ranges in the Northern hemisphere (R 2 = .132, p = .003). This pattern disappeared, however, after two narrow-ranged, but high spore volume taxa in the Humaria genus were removed.

| D ISCUSS I ON
AM fungal taxa had larger distributions than EM fungal taxa. However, within both AM and EM fungal guilds closely related taxa had similar range sizes, suggesting some influence of shared functional trait determination of range sizes. One clear way that closely related fungal taxa may have similar range sizes is if they have similar spore sizes that determine the ability of passive air-borne dispersal. However, spore size was not related to most range size metrics for either mycorrhizal fungal guild. Instead, AM fungi had larger spores and larger ranges than EM fungi. Spore size may not relate to range size if passive wind is not the main dispersal vector for mycorrhizal fungal spores.
Many AM fungi may rely on passive dispersal via seawater (Davison et al., 2018) or active animal dispersal vectors (Mangan & Adler, 2000) and recent evidence suggests that migratory birds may co-disperse AM fungi and seeds across continents (Correia,   sizes compared to mostly wind dispersed EM fungi (Peay et al., 2012, but see Halbwachs, Brandl, & Bassler, 2015. Prolonged passive wind dispersal of smaller, epigeous produced (Kivlin et al., 2014) or melanized (Gessler, Egorova, & Belozerskaya, 2014) spores may explain why spore size was negatively correlated to range size of EM fungi in the Northern hemisphere. Alternatively, EM fungi with smaller spores may have higher UV or desiccation tolerance, allowing for longer air transport (Tipton et al., 2019).
Another phylogenetically conserved trait within mycorrhizal fungal guilds may be the breadth of environmental conditions under which mycorrhizal fungal taxa can persist. In this case, AM fungi may have larger range sizes because they have broader environmental tolerances than EM fungi. AM fungi have been isolated from ecosystems ranging from deserts (Mohammad, Hamad, & Malkawi, 2003) to rainforests (Gaudarrama & Alvarez-Sanchez, 1999), and myriad ecosystems in between (Kivlin, Hawkes, & Treseder, 2011). Moreover, most individual AM fungal taxa have been sequenced from 4 to 5 different ecosystems (Opik et al., 2010). In contrast, EM fungi mostly originate from temperate (Talbot et al., 2014), boreal (Read, Leake, & Perez-Moreno, 2004) or tropical forests (Peay, Baraloto, & Fine, 2009) with some species isolated from in artic and alpine ecosystems (Gardes & Dahlberg, 1996).  Kivlin, Muscarella, Hawkes, and Treseder (2017) found little difference in niche envelopes based on taxonomic resolution. Finally, neither gene region nor DNA similarity influenced community patterns of AM fungal assemblages at the global scale (Kivlin et al., 2011).
Nevertheless, these results should be interpreted with these species definitions in mind.

| CON CLUS ION
Overall, I found little indication that dispersal limitation measured as a proxy of spore size is the main driver of mycorrhizal fungal distributions at the global scale. Instead range sizes were phylogenetically conserved both within AM and EM fungal lineages. This suggests that either other mechanisms of dispersal are occurring or that filtering by abiotic or biotic factors influences mycorrhizal fungal range size dynamics. Future modelling including these environmental imprints on mycorrhizal fungal distributions will elucidate the relative influence of dispersal limitation versus abiotic and biotic filtering on mycorrhizal fungal distributions.

ACK N OWLED G EM ENTS
This work was supported by Texas Ecology Laboratory funding (2018-2020) to SK. Earlier drafts of this manuscript were improved through comments from CV Hawkes and C Averill.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are openly available in DRYAD at https://doi.org/10.5061/dryad.b8gth t78v