{"paper_id":"0df52a18-be4b-4a8a-aaf3-94d7bb00a2e4","body_text":"1 \n \nRoot hairs and mycorrhiza represent alternative phylogenetically conserved 1 \nstrategies for belowground absorptive surface maximization 2 \nJoana Bergmann*1,2, Tom Lachaise3, Karla M. Barfuss1,2, Emma Bretherick4,5, Elsa Matthus1, Mark van 3 \nKleunen3,6, Matthias C. Rillig2,7 4 \n 5 \n1Leibniz Centre for Agricultural Landscape Research (ZALF), Sustainable Grassland Systems, 6 \nEberswalder Straße 84, 15374 Müncheberg, Germany 7 \n2Freie Universität Berlin, Plant Ecology, Altensteinstr 6, 14195 Berlin, Germany 8 \n3Universität Konstanz, Ecology, Universitätsstr 10, 78457 Konstanz, Germany 9 \n4Universidad Nacional Autónoma de México, Instituto de Ecología, Avenida Universidad 3000, 04510 10 \nCoyoacán, Mexico City, Mexico 11 \n5Humboldt-Universität zu Berlin, Unter den Linden 6, 10117 Berlin, Germany 12 \n 6Zhejiang Key laboratory for Restoration of Damaged Coastal Ecosystems & Zhejiang Provincial Key 13 \nLaboratory of Plant Evolutionary Ecology and Conservation, Taizhou University, Taizhou 318000, 14 \nChina 15 \n7Berlin Brandenburg Institute of Biodiversity research (BBIB) 16 \n*corresponding author: Joana Bergmann, joana.bergmann@zalf.de, +49 (0)33237 849 31 17 \n 18 \nMain text body: 5916 words 19 \nSummary: 199 words 20 \nIntroduction: 1063 words 21 \nMaterial and Methods: 2116 words 22 \nResults: 877 words, 5 figures in colour 23 \nDiscussion incl. Conclusion: 1860 words 24 \nSupplementary Information: 4 figures in colour, 5 tables 25 \n26 \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 14, 2026. ; https://doi.org/10.64898/2026.05.13.723781doi: bioRxiv preprint \n\n2 \n \nSummary 27 \n• Plants take up nutrients from the soil while investing in absorptive root surface or 28 \nmycorrhizal partners. Root hairs - a major structure for nutrient uptake and cheap to build - 29 \nincrease the absorptive root surface. As such they are an important component of plant 30 \nresource economics but largely neglected in root economic concepts so far.  31 \n• This is mainly due to data scarcity, which we set out to overcome by measuring root-hair 32 \ntraits on 82 European grassland species in a greenhouse experiment. Using fluorescence and 33 \nlight microscopy, root-hair length and incidence was measured along with mycorrhizal 34 \ncolonization.  35 \n• We found a phylogenetically conserved trade-off between plant investment into root hairs 36 \nand mycorrhiza. A similar trade-off between root-hair incidence and mycorrhiza occurred at 37 \nthe intraspecific level, while patterns were heterogeneous among species. Plant species with 38 \nhigh colonization rates showed the highest variability in root-hair incidence. 39 \n• We conclude that plants vary along a gradient ranging from investment into root hairs as 40 \npart of a “do-it-yourself” strategy to collaboration with mycorrhizal fungi while showing 41 \nintraspecific variation in root-hair incidence. These findings demonstrate that root hairs play 42 \na fundamental role in fine-root trait variation and need to be considered when studying 43 \nbelowground plant economic strategies. 44 \n 45 \nKeywords 46 \nCollaboration, fine roots, outsourcing, root economics space, functional strategy, belowground traits 47 \n  48 \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 14, 2026. ; https://doi.org/10.64898/2026.05.13.723781doi: bioRxiv preprint \n\n3 \n \nIntroduction 49 \nThe resource economy of plants has been a focal area of studies investigating plant functional traits 50 \n(Wright et al., 2004; Freschet et al., 2013a,b; Reich, 2014; Kong et al., 2017; Bergmann et al., 2020; 51 \nWeigelt et al., 2021, Carmona 2021, Matthus 2025). The general idea is that plants invest carbon in 52 \nconstruction and conservation of tissue to ensure the uptake and transport of resources. For 53 \naboveground organs - mainly leaves - an economic spectrum of plant strategies has been described 54 \nand confirmed on a global basis (Wright et al., 2004; Reich, 2014; Díaz et al., 2016). This spectrum 55 \nranges from fast growth and resource acquisition of short-lived organs to slow but steady resource 56 \nacquisition of organs constructed for longevity. An analogue pattern is found in belowground fine-57 \nroots and called the conservation gradient (Bergmann et al., 2020; Weigelt et al., 2021; Matthus et 58 \nal., 2025).  59 \nThe concept of a collaboration gradient in root-trait variation that is independent of the conservation 60 \ngradient and unique to belowground economy has been proposed (Bergmann et al., 2020) and found 61 \nto be a solid pattern across organizational levels and study systems (Matthus et al., 2025). This 62 \ncollaboration gradient describes plant strategies in soil exploration ranging gradually from do-it-63 \nyourself investment in specific root length (SRL) to outsourcing to mycorrhizal fungal partners with 64 \nthe consequence of larger fine root diameter (AD) and cortex fraction (CF).  65 \nArbuscular mycorrhizal (AM) fungi, which associate with almost 80% of all land plants (Brundrett & 66 \nTedersoo, 2018), colonize the root’s cortex and explore the soil with extraradical hyphae (Smith & 67 \nRead, 2008). AM fungi can take up limiting nutrients like phosphorus and nitrogen, supplying them to 68 \nthe roots in exchange for carbon synthesized by the plant partner´s aboveground photosynthesis 69 \n(Bolan et al., 1987; Smith & Read, 2008). Besides the exploration of a certain volume of soil, the 70 \nactual surface and the soil contact of an absorptive plant or fungal structure determines the rate of 71 \nreturn on investment of a plant (McCormack & Iversen, 2019). Little is known about traits of fungal 72 \nextraradical hyphae, but a large body of literature reveals that for the plant itself an effective way to 73 \nmaximize a bsorptive surface and soil exploration is the production of root hairs (Bhat & Nye, 1973; 74 \nGahoonia et al., 1997; Bates & Lynch, 2000a,b; Haling et al., 2013; Brown et al., 2013b). Yet, most 75 \nlikely because of the effort involved in data collection, the coverage of root-hair traits in databases is 76 \npoor (Iversen et al., 2017; Guerrero-Ramirez, 2020; Kattge et al., 2020), and their integration into 77 \nbroader plant economics concepts is inconclusive  (Zhao et al., 2024; Matthus et al., 2025). 78 \nRoot hairs are unicellular epidermal extensions on living fine roots of most land plants (Farquhar, 79 \n1996). They enhance nutrient and water uptake of fine roots (Bhat & Nye, 1973; Gilroy & Jones, 80 \n2000; Haling et al., 2013; Carminati et al., 2017; Freschet et al., 2021b) contributing to >60% of the 81 \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 14, 2026. ; https://doi.org/10.64898/2026.05.13.723781doi: bioRxiv preprint \n\n4 \n \nplant´s phosphorus-demand (Gahoonia & Nielsen, 1998). Root-hair traits are known to widely vary 82 \namong species and along environmental gradients of soil fertility (Lambers et al., 2008; Holdaway et 83 \nal., 2011), while a large root-hair surface can be realized with long (Yang et al., 2015; Haling et al., 84 \n2016) and/or many (Brown et al., 2013a; Marzec et al., 2015) root hairs. Furthermore, they have 85 \nbeen described to be comparably responsive to soil P availability, in part because of their dynamic 86 \ngrowth and life-span (Bates & Lynch, 1996; Zhu et al., 2010; Nestler & Wissuwa, 2016). Carbon-87 \ncheap, metabolically active (Ma et al., 2018) and dynamic in construction compared to fine roots, 88 \nroot hairs might therefore resemble a ‘fast’ economic strategy component within the conservation 89 \ngradient. So far, no study could provide empirical data to support this link (Matthus et al., 2025).  90 \nEmpirical evidence suggests that the plant species specific beneficial effect of being mycorrhizal is 91 \nrelated to root-hair length (Bolan et al., 1987; Schweiger et al., 1995). Therefore, it has been 92 \nhypothesized that the carbon investment into root hairs might be an alternative strategy to the 93 \nmycorrhizal symbiosis on an interspecific (Maherali, 2017) and intraspecific level (Kumar et al., 2019) 94 \nfor acquiring soil resources. If this pattern were to be verified for a larger set of species, it would 95 \nimply that root hairs represent another aspect of a do-it-yourself strategy of plant economics within 96 \nthe collaboration gradient. To date, this hypothesis has not been tested for a larger species set. For 97 \nfour monocotylous families, Betekhtina et al. (2023) provide evidence for root hair length to trade-98 \noff with AMF colonization levels. Contrary, Guilbeault-Mayers et al. (2024) reported a root-hair index 99 \n(encompassing length and density) to increase with AMF colonization in trees. Parasquive et al. 100 \n(2023) found opposing intraspecific trends of root hair index loadings along the collaboration 101 \ngradient, depending on tree species. Another study found root hair traits to be independent of the 102 \ncollaboration gradient (Guilbeault-Mayers et al., 2024), as has also been proposed by Dallstream & 103 \nSoper (2024). 104 \nIt has long been known that non-mycorrhizal plants typically have many and long root hairs while 105 \nmycorrhizal plants often lack them ( Kelley, 1950; Schweiger et al., 1995; Jakobsen et al., 2005; 106 \nBrundrett, 2021). Most of these studies only worked with mycorrhizal status as a categorical 107 \nclassification to classify plant functioning. Brundrett & Tedersoo (2018) already noted though, that 108 \nplant species with many root hairs normally have low mycorrhization and can typically be found in 109 \nstressful habitats. Still, the current data available originate from different studies conducted under 110 \nvarious conditions and measuring different traits and categories, which makes it hard to test for 111 \ngradual functional trade-offs. 112 \nTo fill this knowledge gap, we measured root-hair length (HL) and root-hair incidence (HI) as well as 113 \nmycorrhizal colonization on a large set of grassland species grown under common conditions. We 1) 114 \ntested for phylogenetic patterns and differences between functional groups and mycorrhizal status. 115 \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 14, 2026. ; https://doi.org/10.64898/2026.05.13.723781doi: bioRxiv preprint \n\n5 \n \nWe were interested in whether root hair traits differ between these widely used categories, or 116 \nwhether they gradually change with mycorrhizal colonization. We further aimed to 2) test the 117 \nhypothesis of an interspecific trade-off between the investment in root hairs and the mycorrhizal 118 \npartner and to 3) explore the intraspecific variation of root hair traits. Finally, we aimed to 4) 119 \nintegrate interspecific variation of HL and HI into the concept of the root economics space, 120 \nhypothesizing that root hair investment represents a do-it-yourself strategy additional to the overall 121 \nincrease of SRL. 122 \n 123 \n  124 \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 14, 2026. ; https://doi.org/10.64898/2026.05.13.723781doi: bioRxiv preprint \n\n6 \n \nMaterial and Methods 125 \n 126 \nSpecies set 127 \nThe experiment was conducted in the framework of the Biodiversity Exploratories (Fischer et al., 128 \n2010), a large scale and long term land-use experiment with 150 grassland plots located in three 129 \nareas in northern, central and southern Germany. From the vegetation records of the Exploratories, 130 \nwe chose a set of 94 grassland species which could be purchased from the commercial seed supplier 131 \nRieger-Hofmann GmbH (Blaufelden-Raboldshausen, Germany). This species set encompasses 132 \nFabaceae (legumes), non-leguminous dicotyledons (subsequently called forbs) and monocotyledons 133 \n(subsequently called grasses, but note that Allium schoenoprasum is attached here as a 134 \nmonocotyledonous plant). 135 \n 136 \nGreenhouse experiment 137 \nAll data presented here originate from one pot experiment conducted under controlled greenhouse 138 \nconditions, at the facilities of Freie Universität Berlin, between February and June 2018 (16 h day at 139 \n22°C, 8 h night at 15°C). We set up the experiment with 94 initial species, two treatments (with and 140 \nwithout mycorrhizal inoculation), and 8 replicates distributed over 4 overlapping time blocks of 6 141 \nweeks growing time each. The entire experiment therefore consisted of 4 time blocks x 94 species x 2 142 \ntreatments x 2 replicates = 1504 experimental units. Whenever a replicate did not survive, we tried 143 \nto substitute it in the next time block. Nevertheless, some species did not reach a replication of 8 per 144 \ntreatment and some did not germinate at all. In the final analysis, we only included species with a 145 \nminimum of 3 successful replicates per treatment leading to a total of 1151 experimental units of 82 146 \nspecies from 20 families. 147 \nPrior to the first time block, all seeds were surface sterilized in paper tea bags for 3 min in 7% bleach 148 \nfollowed by washing in de-ionized (DI) water until the smell of bleach was gone. The seeds were 149 \ndried at 20°C and stored until sowing for subsequent time blocks. Seeds germinated in plastic boxes 150 \nfilled with 1:1 steamed sand and vermiculite (1-3 mm, ISOLA Vermiculite GmbH; Sprockhövel, 151 \nGermany). Based on germination times recorded in pre-experiments, we sowed the seeds to assure 152 \nthat all seedlings were in the cotyledon stage or had their first leaves developed at time of 153 \ntransplanting. Seedlings were transplanted into plastic cones (410 ml 0.41 L; Stueve & Sons; USA) 154 \nfilled with the same substrate as for germination.  155 \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 14, 2026. ; https://doi.org/10.64898/2026.05.13.723781doi: bioRxiv preprint \n\n7 \n \nThe mycorrhizal treatment was realized as follows: after filling the cone to c. ¾ we added a 30 ml 156 \nhorizon of a 1:1 mixture of steamed sand and mycorrhizal inoculum in vermiculite (INOQ Agri, Inoq 157 \nGmbH, Schnega, Germany). According to the supplier, the inoculum contains 145 spores/ml of 158 \nRhizophagus irregularis propagated on vermiculite (1-2 mm) under non-sterile greenhouse 159 \nconditions. Rhizophagus irregularis is a generalist AM fungus associating with almost all mycorrhizal 160 \nplants (van der Heijden et al., 2015). To account for other soil microbes that might be present in AM 161 \ninoculum produced under non-sterile conditions, we prepared a microbial wash from the 162 \nRhizophagus inoculum (20 µm mesh, soil:water-ratio: 1:2) for the non-mycorrhizal treatment. We 163 \ncarefully adjusted the amount of inoculum as well as the amount of DI water used to prepare the 164 \nmicrobial wash to make sure that each pot received the approximate same number of microbial units 165 \nirrespective of the treatment. To control for nutrients and physical structure of the AM inoculum, we 166 \nautoclaved the solid inoculum used to prepare the microbial wash and added a 1:1 mixture with 167 \nsteamed sand as a horizon to the non-mycorrhizal treatment. For both treatments the added 168 \nhorizons were covered with another layer of ~30 ml substrate to avoid cross contamination. During 169 \ntransplanting, seedlings of the non-mycorrhizal treatment received 30 ml of the microbial wash, 170 \nwhile seedlings of the mycorrhizal treatment received 30 ml of DI water. We replaced seedlings that 171 \ndied shortly after transplanting during the first week.  172 \nWithin each time block, plants grew for 6 weeks in the cones before harvest. All cones were fully 173 \nrandomized at time of transplanting and were rearranged every two weeks. Plants received 25 ml of 174 \nDI water 3 times a week; two weeks and four weeks after transplanting, they received 25 ml of a ¼ 175 \nstrength Hoagland solution (recipe available in Lachaise et al., 2021) instead. 176 \nAt time of harvest, aboveground and belowground biomass of the plants were separated. Roots were 177 \nfirst rinsed with water. Three first order roots per plant were carefully cut, transferred to 10% 178 \nformalin at pH 7 (ROTI Histofix, Carl Roth, Karlsruhe, Germany) in Phosphate Buffered Saline (PBS) 179 \nbuffer and kept at 4 °C for overnight fixation. The next day, the formalin solution was first replaced 180 \nby PBS buffer twice for approx. 2 hours each and finally by a solution of 70% ethanol, 5% glycerin and 181 \n25% DI water for long term preservation. The remaining roots were carefully washed by hand during 182 \nharvest, transferred to cold DI water and kept at 4°C. Within a week they were scanned in water-183 \nfille\nd plastic trays using an Epson perfection 800 Photo scanner at a resolution of 800 dpi. As root 184 \nsystems were small and young, we decided to measure traits on the entire root system. This included 185 \nmainly first to third order roots and a small fraction of higher order roots. 99.98 % of root length 186 \nwithin the entire experiment belonged to roots with a diameter < 2 mm.  187 \n  188 \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 14, 2026. ; https://doi.org/10.64898/2026.05.13.723781doi: bioRxiv preprint \n\n8 \n \nTrait measurements 189 \nTotal root length and volume as well as the average root diameter (D [mm]) were measured using 190 \nWinRhizo 2017 software (Regent Instruments Inc., Québec, Canada). Aboveground and belowground 191 \ndry biomass was determined after drying at 60°C for at least three days. The mycorrhizal growth 192 \nresponse (MGR) of each plant species was calculated as MGR= ln[total dry biomass inoculated / total 193 \ndry biomass non-inoculated] (Hoeksema et al., 2010; Maherali, 2014). All other root traits were 194 \nmeasured within the mycorrhizal treatment assuming that this resembles the natural soil biotic 195 \ncondition. Root dry biomass was used to calculate the specific root length (SRL – root length/dry 196 \nbiomass [m/g]) and root tissue density (RTD – dry biomass/root volume [g/cm³]) by calculating the 197 \nvolume as the sum of 0.2 mm diameter size classes according to Rose (2017).  198 \nRoot-hair length (HL [µm]), cortex fraction (CF [%]) and first order root diameter (Dfirst [mm]) were 199 \nmeasured on the preserved first order root tips of three randomly chosen replicates per species from 200 \nthe mycorrhizal treatment using a fluorescence microscope (Zeiss Axio Imager 2, Carl Zeiss AG, 201 \nOberkochen, Germany). One root per replicate was randomly picked and dyed in 0.01% Calcofluor-202 \nwhite (Thermo Fisher Scientific, Waltham, USA) for 5-10 seconds. Subsequently, it was rinsed in DI 203 \nwater for a few seconds, mounted on a slide and carefully covered with a cover slip without applying 204 \npressure. As Calcofluor-white binds to cellulose, it helps distinguish plant cell walls including those of 205 \nfine root hairs. As all roots were small and translucent there was no need for cross sectioning to 206 \nmeasure stele and cortex diameter (Fig. S1). Microscopic images were taken with a Zeiss AxioCam at 207 \na magnification of x50 using a 430 nm fluorescence filter. For each replicate, several images were 208 \ntaken using the functions “Z-Stacks” and “Tiles” to display a continuous segment of 5 mm within the 209 \nmature root hair zone. The “Tiles” function merges several images along the root while the “Z-210 \nStacks” function combines images vertically, thereby producing an in-focus image throughout the 211 \nentire range of depths. We defined the Z-Stacks to range from the middle of the stele to the upper 212 \nepidermal layer of the root just beneath the cover slip. The first order root diameter (Dfirst) as well as 213 \nthe stele diameter were measured at three positions along the image. We calculated the cortex 214 \nfrac\ntion (CF) as the percent area of a first order root cross section that is occupied by tissue outside 215 \nthe stele (Freschet et al., 2021a). Mean values of Dfirst and CF were first calculated at replicate level 216 \nand subsequently at the species level. HL was measured according to Delhaize et al. (2012). In brief, 217 \nwe divided the 5 mm root segments into 5 sub-segments of 1 mm each and measured the length of 218 \nthe longest root hair on each side of the root in each sub-segment (i.e. 2 root hairs per sub-segment). 219 \nAll 10 measurements per 5 mm root segment were averaged to calculate the mean HL per replicate.  220 \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 14, 2026. ; https://doi.org/10.64898/2026.05.13.723781doi: bioRxiv preprint \n\n9 \n \nMycorrhizal status (obligate mycorrhizal, facultative mycorrhizal, non-mycorrhizal) was assigned on 221 \nspecies level according to the FungalRoot database (Soudzilovskaia et al., 2020). In case of conflicting 222 \nstatus reports within the database, we followed the provided expert recommendations. 223 \n 224 \nFor the determination of the percentage of mycorrhizal colonization (%M) and the root-hair 225 \nincidence (HI [%]), we used representative subsamples of the dried root systems of the three 226 \nreplicates from the mycorrhizal treatment of each species (Freschet et al., 2021a). Furthermore, one 227 \nreplicate per species in the non-mycorrhizal treatment was randomly chosen and checked for AM 228 \ncolonization. Roots were first cleared in 10% (w/v) KOH for 15 min at 80°C and then stained in 0.05% 229 \n(w/v) Trypan Blue in lactoglycerol for another 15 min at 80°C. Mycorrhizal colonization was 230 \ndetermined with the magnified intersection method (McGonigle et al., 1990) at a magnification of 231 \nx200, using a minimum of 30 root pieces on a slide to count presence or absence of mycorrhizal 232 \nstructures in general (colonization rate) and of arbuscules in specific (rate of arbuscular colonization) 233 \nin 50-100 intersects. Mycorrhizal hyphae were identified based on their staining, their growing habit 234 \n(intraradical between cortical cells) and their missing of irregular septation. Due to the commercial 235 \ninoculum, there was very little contamination with other fungi. For the non-mycorrhizal treatment, 236 \nmycorrhizal colonization rates between 1% and 6% were detected for 6 out of 82 replicates 237 \nsuggesting very limited contamination. Within the mycorrhizal treatment colonization rates of up to 238 \n86% and rates of arbuscular colonization up to 72% clearly confirmed a successful inoculation. HI was 239 \ndetermined simultaneously and analogously to %M, recorded as presence or absence at each 240 \nintersect (Siqueira & Saggin-Júnior, 2001), giving a proxy of how much percent of the root length was 241 \ncovered by root hairs.  242 \nThe coefficients of variation of HI and HL (cvHI, cvHL) were calculated by using the general R function 243 \ncv(x) that computes the sample coefficient of variation as (SD/mean)*100. To display the within-244 \nspecies correlation between %M and HI of all species while accounting for overall between-species 245 \ndifferences in both traits, we normalized the data by coding the intraspecific median of the three 246 \ntrait records per species as 0, the lower value as = lower value – median value and the higher value 247 \nas = h igher value – median value. 248 \nRoot nitrogen concentration (N [%]) was measured on the three replicates, after drying and milling 249 \nthe roots, using an Elemental Analyzer (Euro EA, HEKAtech, Wegberg, Germany).  250 \n 251 \n  252 \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 14, 2026. ; https://doi.org/10.64898/2026.05.13.723781doi: bioRxiv preprint \n\n10 \n \nAnalysis 253 \nAll analyses were carried out in R version 3.6.3 (R Core Team, 2020). To explore phylogenetic 254 \npatterns in the root hair data we used the function drop.tip() from the package ape (Paradis et al., 255 \n2004) to prune the DaPhnE phylogeny (Durka & Michalski, 2012) for our species set and the function 256 \nphylosig() as well as phylo.heatmap() from the package phytools (Revell, 2012) to calculate the 257 \nphylogenetic signal of all traits and to display trait variation along the tree by using color palettes 258 \nfrom the package viridis (Garnier, 2018). The package ggplot2 (Wickham, 2010) was used to display 259 \nviolin plots, the pairwise correlation heatmap and the correlation between %M and HI at intraspecific 260 \nlevel and the package cowplot (Wilke, 2024) was used for multipanel figures. 261 \nPrior to the calculation of pairwise correlations and principal component analyses (PCA), we 262 \nimproved data distribution by applying log transformation for all traits except CF, HI and %M, which 263 \nwe transformed using the function logit() from the gtools (Warnes et al., 2020) package, since these 264 \ntraits varied between 0 and 1. The function rcorr() from the package Hmisc (Harrell, 2020) was used 265 \nto calculate Pearson’s correlations of all trait pairs, and the functions comparative.data() and pgls() 266 \nfrom the package caper (Orme et al., 2018) were used to calculate phylogenetically corrected 267 \npairwise correlations. We determined the phylogenetically corrected correlation coefficient by taking 268 \nthe square root from the adjusted r² of the model multiplied by -1 in case of a negative regression 269 \ncoefficient while assigning r=0 in case of negative adjusted r² values. The phylogenetically informed 270 \nPCAs were calculated using phyl.pca() from the package phytools (Revell, 2012) and displayed using 271 \nfunctions from the package shape (Soetaert, 2014). Ellipsoids were plotted using the package ellipse 272 \n(Murdoch & Chow, 2024). Permanova based on euclidean pairwise distances in PCA space among 273 \ngroups were performed using the pairwise.adonis() function from the package pairwiseAdonis 274 \n(Martinez Arbizu, 2019). 275 \nWe further used the packages dplyr (Wickham et al., 2020a), Rmisc (Hope, 2013), raster (Hijmans, 276 \n2020), data.table (Dowle & Srinivasan, 2020) and devtools (Wickham et al., 2020b) for general data 277 \nhandling and exploration. 278 \n 279 \nResults 280 \nRoot hair traits show a phylogenetically conserved pattern 281 \nR\noot-hair length and incidence showed strong phylogenetic signals (Fig. 1, Table S1) and were highly 282 \npositively correlated (Fig. S2). Monocotyledons showed many long root hairs, with the hairless Allium 283 \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 14, 2026. ; https://doi.org/10.64898/2026.05.13.723781doi: bioRxiv preprint \n\n11 \n \nschoenoprasum being the only exception, while legumes (Fabaceae) had few and short root hairs 284 \n(Fig. 1). Within the other dicotyledonous families, Asteraceae showed low values while 285 \nPolygonaceae, Caryophyllaceae and Brassicaceae showed high values for both root-hair length and 286 \nincidence. Throughout the entire set of species, mycorrhizal colonization showed a completely 287 \ninverted pattern which was strongly phylogenetically conserved as well (Fig. 1, Table S1). Clades with 288 \nlong root hairs and high root-hair incidence were poorly colonized by mycorrhizal fungi, while clades 289 \nwith short and few root hairs showed high colonization rates. Root-hair length and incidence were 290 \nboth negatively correlated with mycorrhizal colonization. This pattern disappeared for root-hair 291 \nlength after phylogenetic correction (Fig. S2). 292 \nExamining plant functional types, root-hair length was lower in legumes than in grasses and forbs, 293 \nwith grasses having the longest root hairs overall (Fig. 2a). The coefficient of variation in root-hair 294 \nlength did not differ significantly among these plant functional groups, even though grasses tended 295 \nto have the least variation (Fig. 2e). Root-hair incidence was higher and its coefficient of variation 296 \nwas lower in grasses than in both legumes and forbs (Fig. 2b,f). 297 \n 298 \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 14, 2026. ; https://doi.org/10.64898/2026.05.13.723781doi: bioRxiv preprint \n\n12 \n \n 299 \nFig. 1: The phylogenetically conserved trade-off between the investment in root hairs and 300 \nmycorrhization. Colour-coded are the species mean values of root-hair length and incidence as well 301 \nas percent mycorrhizal colonization to the right and the corresponding phylogenetic tree with 302 \nbroader taxonomic groups to the left. Trait values are standardized to the same range, colour-coded 303 \nfrom yellow (low) via green (medium) to blue (high). Phylogenetic signal of each trait is displayed as 304 \nPagel´s lambda. Grasses are shaded in medium grey, non-leguminous forbs in light grey and 305 \nleguminous forbs in dark grey. Allium schoenoprasum L. with the lowest hair incidence had no root 306 \nhairs in the samples for determination of hair length leading to a missing value. 307 \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 14, 2026. ; https://doi.org/10.64898/2026.05.13.723781doi: bioRxiv preprint \n\n13 \n \n 308 \nFig. 2: Variation in root hair traits according to plant functional type and mycorrhizal status. 309 \nDisplayed are raw data (upper panels) of species mean root-hair length (HL, panel a, c) and incidence 310 \n(HI, panel b, d) as well as the coefficient of variation (lower panels) in root-hair length (cvHL, panel e, 311 \ng) and incidence (cvHI, panel f, h). Displayed are kernel density distributions and group means (black 312 \ndots) with 95% confidence intervals. Non-overlapping confidence intervals are highlighted by a 313 \ncoloured ribbon to visualize group differences. Plant functional types: grasses, forbs, legumes; 314 \nmycorrhizal status: obligate mycorrhizal (AM), facultative mycorrhizal (AM-NM), non-mycorrhizal 315 \n(NM). 316 \n 317 \nMycorrhizal status is not a strong predictor of root-hair traits 318 \nOf the 82 grassland species tested in the experiment, 64 were obligate mycorrhizal, 9 facultative 319 \nmycorrhizal and 9 non-mycorrhizal, as classified according to the FungalRoot database. Mycorrhizal 320 \nstatus did not predict species root-hair length or its coefficient of variation well, even though 321 \nobligate mycorrhizal species tended to have shorter root hairs (Fig. 2c). Root-hair incidence instead 322 \nwas lower in obligate mycorrhizal species than in non-mycorrhizal species, while facultative 323 \nmycorrhizal species had intermediate values (Fig. 2d). The coefficient of variation of root-hair 324 \nincidence showed the opposite pattern with obligate mycorrhizal species being more variable than 325 \nnon-mycorrhizal species, and facultative mycorrhizal species showing intermediate values again (Fig. 326 \n2h). 327 \n  328 \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 14, 2026. ; https://doi.org/10.64898/2026.05.13.723781doi: bioRxiv preprint \n\n14 \n \nAn ecological trade-off between root hairs and mycorrhization 329 \nThe phylogenetically informed PCA revealed a strong trade-off (PC 1 = 45%) between high root-hair 330 \nincidence and length on one end, and high mycorrhizal colonization rates on the other, accompanied 331 \nby an increase in variation in root-hair incidence (Fig. 3, Table S2). PC 2 explained 23% of variation, 332 \nwith the coefficient of variation of root-hair length influencing this axis most strongly.  333 \nMycorrhizal status as well as plant functional types affected species locations within the PCA (Table 334 \nS3). Non-mycorrhizal species differed from obligate mycorrhizal species by being closely aggregated 335 \nat high values of root-hair incidence on PC 1. Grasses differed from both forbs and legumes by 336 \nshowing high root-hair incidence as well, even though considerable variation occurred within each of 337 \nthe functional types. Legumes were located at high values of mycorrhizal colonization and variation 338 \nin root-hair incidence on PC 1 while spanning the entire range of PC 2. 339 \n 340 \n 341 \n 342 \nFig. 3: Phylogenetically informed principal component analysis of root hair traits and mycorrhizal 343 \ncolonization rate. Panel a displays species based on their mycorrhizal status (obligate mycorrhizal - 344 \nAM, facultative mycorrhizal - AM-NM, non-mycorrhizal - NM) while panel b displays species based on 345 \ntheir plant functional group (grasses, forbs, legumes). Ellipsoids and large dots display 95% 346 \nconfidence intervals and centroids. PCA results can be found in table S2. HL – hair length, HI – hair 347 \nincidence, cvHL – coefficient of variation in hair length, cvHI – coefficient of variation in hair 348 \nincidence, %M - percent mycorrhizal colonization. 349 \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 14, 2026. ; https://doi.org/10.64898/2026.05.13.723781doi: bioRxiv preprint \n\n15 \n \n 350 \nAt the intraspecific level, root-hair incidence correlates with mycorrhizal colonization rate but 351 \nspecies show strong heterogeneity  352 \nDespite low within-species replication (n = 3 individual plants per species scored for root hair traits), 353 \nwe could detect an overall association between root-hair incidence and mycorrhizal colonization rate 354 \n(Fig. 4). Individual plants with higher colonization rates had lower root-hair incidence than less-355 \ncolonized individuals within the same species. Overall, we found a slight negative correlation (slope = 356 \n-0.29, p<0.001) with small confidence intervals. However, there was considerable variation among 357 \nspecies. No intraspecific correlation was found between the mycorrhizal colonization rate and root-358 \nhair length. 359 \n 360 \n 361 \nFig. 4: Intraspecific correlation of root-hair incidence and mycorrhizal colonization. Displayed is the 362 \nrelative difference in mycorrhizal colonization (%M) and root-hair incidence (HI) within each species 363 \nas well as the overall correlation with 95% confidence interval. 364 \n 365 \nRoot-hair traits add to the root economics space  366 \nThe inclusion of root-hair traits introduced a new dimension to the root economics space. The first 367 \naxis (22%) of the extended PCA (Fig. 5, Fig. S3, Table S4) was dominated by specific root length on 368 \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 14, 2026. ; https://doi.org/10.64898/2026.05.13.723781doi: bioRxiv preprint \n\n16 \n \none end and root average diameter and diameter of first order roots on the other, accompanied by 369 \nvariation in cortex fraction but also root-tissue density. The trade-off between the investment in 370 \nroot-hair length and incidence on one end and the coefficient of variation of root-hair incidence on 371 \nthe other dominated the second axis (20%). Mycorrhizal colonization intensity and mycorrhizal 372 \ngrowth response were associated with variation in root-hair incidence. Root-tissue density loaded 373 \nstrongest on the third axis (14%) together with the coefficient of variation of root-hair length and 374 \nantagonistically to root-nitrogen concentration and cortex fraction.  375 \nPlants of different mycorrhizal status as well as different plant functional types differed in their root 376 \neconomic strategies within the space (Table S5). Non-mycorrhizal and facultative mycorrhizal plants 377 \ndiffered from obligate mycorrhizal plants, while non-mycorrhizal plants showed the highest specific 378 \nroot length on PC1 and highest root-hair length and incidence on PC2, and there was no general 379 \npattern along PC3. Obligate mycorrhizal plants spanned the entire space but clearly showed the 380 \nhighest values for root diameter on PC1, lowest root-hair incidence and highest colonization rate on 381 \nPC2 and highest root-nitrogen concentration on PC3. Legumes were located at high root diameter, 382 \ncortex fraction and colonization rate as well as high root-nitrogen concentration. Grasses showed a 383 \nclear trend towards high root-hair length and incidence as well as specific root length. As such, 384 \ngrasses and legumes formed distinct groups almost without overlap, while forbs spanned the entire 385 \nroot economics space. 386 \n 387 \n 388 \n 389 \nFig. 5: Extended phylogenetically informed principal component analysis. Displayed are species 390 \nbased on their mycorrhizal type (obligate mycorrhizal - AM, facultative mycorrhizal – AM-NM, non-391 \nmycorrhizal - NM). Ellipsoids and large dots display 95% confidence intervals and centroids. PCA 392 \nresults can be found in Table S4. HL – hair length, HI – hair incidence, cvHL – coefficient of variation 393 \nin hair length, cvHI – coefficient of variation in hair incidence, %M - % mycorrhizal colonization, SRL – 394 \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 14, 2026. ; https://doi.org/10.64898/2026.05.13.723781doi: bioRxiv preprint \n\n17 \n \nspecific root length, AD – average diameter, Dfirst – diameter of first order roots, CF – cortex fraction, 395 \nRTD – root tissue density, N – root nitrogen concentration, MGR – mycorrhizal growth response. 396 \n 397 \nDiscussion 398 \nPlants can either invest in root hairs or rely on mycorrhizal partners while maintaining variation in 399 \nroot-hair incidence 400 \nWe found a striking pattern of an evolutionarily conserved trade-off between plant investment in 401 \nroot hairs – specifically their incidence – and mycorrhizal symbiosis. The phylogenetic conservation 402 \noccurred at high taxonomic levels; grasses showing high root-hair incidence and length, paired with 403 \nlow mycorrhizal colonization rates, while legumes exhibited the opposite pattern (Fig. 1 and 2). This 404 \nsupports existing knowledge on grasses and legumes (Hill et al., 2006). Non-leguminous forbs 405 \nexhibited a range of strategies along the entire gradient of variation. This is an expectable result, 406 \ngiven that forbs are not monophyletic and comprise all forms of mycorrhizal status. An evolutionarily 407 \ndeep-rooted phylogenetic signal might be the reason why many correlations between raw trait data 408 \ndisappeared after phylogenetic correction. 409 \nIn contrast to the level of mycorrhizal colonization, we found mycorrhizal status to be a weak 410 \npredictor of root-hair traits. The traditional mycorrhizal status classification of species as being 411 \nobligate, facultative or non-mycorrhizal and the respective definitions have been discussed lately 412 \n(Cosme et al., 2018; Brundrett & Tedersoo, 2019). Cosme et al. (2018) argue that species classified as 413 \nbeing non-mycorrhizal can have low levels of colonization and even a few arbuscules. We found the 414 \nsame pattern in our species classified as non-mycorrhizal with low colonization rates and no or very 415 \nfew arbuscules. Moreover, we found that these species had high root-hair length and incidence, 416 \nhence resembling an extreme do-it-yourself trait syndrome.  417 \nNon-mycorrhizal plants can be subdivided based on their phosphorus (P) acquisition strategy as P-418 \nscavengers, which rely on dissolved P, and P-miners, which exude organic compounds that release 419 \nfixed P (Lambers et al., 2008; Lambers & Teste, 2013; Yu et al., 2020). Carex vulpina, as the only non-420 \nmycorrhizal species with a P-mining strategy in our dataset showed the highest HI but only an 421 \nav\nerage HL within the non-mycorrhizal status. A larger number of Proteaceae type species (P-miners) 422 \nwould be needed to draw general conclusions about HL and HI patterns of the different non-423 \nmycorrhizal P acquisitions strategies. Hence, our results can be considered representative only for P 424 \nscavenging species.  425 \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 14, 2026. ; https://doi.org/10.64898/2026.05.13.723781doi: bioRxiv preprint \n\n18 \n \nThe classification of facultative mycorrhizal species includes both species that are mycorrhizal only 426 \nunder specific circumstances (like nutrient deficiencies) and species that always have low 427 \ncolonization rates; hence plants with different ecological strategies (Brundrett & Tedersoo, 2019; 428 \nSoudzilovskaia et al., 2020). Furthermore, the status ‘facultative’ can be misleading in case of 429 \nconflicting observations for species with an overall low data record. Accordingly, we found strong 430 \noverlap in root-hair traits between obligate and facultative mycorrhizal species, even though the 431 \nlatter tended to have more and longer root hairs as we would have expected given the fact that they 432 \nhave lower colonization rates (Fig. S4). The categories of facultative and obligate mycorrhizal species 433 \nseem to not be informative for root hair patterns. However, non-mycorrhizal plants differed strongly 434 \nfrom obligate mycorrhizal plants by having higher root-hair incidence and lower variation therein. 435 \nThis pattern also dominated the overall gradual trade-off between the investment in root hairs and 436 \nmycorrhiza: a strong investment in root-hair incidence was accompanied by low variation of the 437 \nsame trait. Species with high mycorrhizal colonization rates produce fewer root hairs but encompass 438 \nmore intraspecific variation.  The coefficient of variation provided us with a scale-independent 439 \nmeasure of variation. It should be noted though that at a given standard deviation, it is inversely 440 \nrelated to the mean value, hence mathematically favoring high mean trait values to coincide with low 441 \nvariation of the same trait. Given the fact that being mycorrhizal can be a competitive advantage in 442 \nmany though not all terrestrial habitats (Brundrett & Tedersoo, 2018), it remains to be studied how 443 \nroot hair traits add to the filtering of environmental variation for species occurrence (Laughlin et al., 444 \n2021). 445 \n 446 \nIntraspecific variation in root-hair incidence and mycorrhizal colonization mirrors the interspecific 447 \npattern 448 \nAlthough this experiment was not designed to test for intraspecific variation, we could show that 449 \noverall, the within-species root-hair incidence was higher at lower colonization rates. We cannot 450 \ndetermine if this variation originates from a plastic response of the plant to different colonization 451 \nlevels of the AM fungus or from genetic variation between plant individuals. Further research is 452 \nneeded to evaluate this question and to determine cause and effect. Plasticity in both root-hair 453 \nlength and incidence has been reported in response to soil P (Bates & Lynch, 1996, 2000b; Zhu et al., 454 \n2010) as well as mycorrhizal inoculation (Price et al., 1989; Sun & Tang, 2013; Wu et al., 2016). 455 \nSuggestions about the resource costs of root hairs being higher (Price et al., 1989) or lower (Brown et 456 \nal., 2013a) than those of the mycorrhizal symbiosis differ widely, while soil moisture and P availability 457 \n(Brown e t al. , 2013a; Fort et al., 2015; Ma et al., 2021) further mediate the effects. Given the design 458 \nof our study with homogeneous soil fertilization, we cannot test the effect of soil P on root hair traits 459 \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 14, 2026. ; https://doi.org/10.64898/2026.05.13.723781doi: bioRxiv preprint \n\n19 \n \nand their variability. Specifically, it remains to be studied if facultative mycorrhizal species show 460 \nhigher coefficients of variation in root-hair incidence and length under different levels of soil P. 461 \nFurthermore, patterns might change over time, given the fact that we analyzed young plants. 462 \nNevertheless, our results suggest an overall intraspecific trade-off between root-hair incidence and 463 \nmycorrhizal colonization mirroring the interspecific pattern and leading to stronger variation in root-464 \nhair incidence in obligate mycorrhizal species with high colonization rates.  465 \n 466 \nRoot hairs add to the do-it-yourself strategy of plants 467 \nThe trade-off between root hairs and mycorrhizal colonization rate defined the second axis of the 468 \nprincipal component analysis on all traits, with the root traits of the collaboration gradient 469 \ndominating the first and those of the conservation gradient the third axis. The first axis resembled 470 \nthe collaboration gradient with a trade-off between ‘do-it-yourself’ with high SRL and ‘outsourcing’ 471 \nwith high root diameter and cortex fraction as expected within the framework of the root economics 472 \nspace (Bergmann et al., 2020; Ding et al., 2020; Wen et al., 2022). RTD also loaded on axis 1 – though 473 \nless than on axis 3 - with a considerable amount of variation, being negatively correlated to SRL. This 474 \ncorrelation has been reported before (Eissenstat, 1992; Reich, 2014) and might originate from the 475 \nfact that, for a given diameter SRL, has to increase with decreasing RTD (Ostonen et al., 2007). This 476 \nmight be the most important driver behind former detection of a one dimensional root economics 477 \nspectrum that parallels leaf economics (Freschet et al., 2010; Reich, 2014).  478 \nOn axis 2, root-hair length and incidence behaved antagonistically to the degree of variation in root-479 \nhair incidence accompanied by mycorrhizal colonization rate and growth response. With 20% 480 \nvariance, the trade-off explained a considerable amount of variation within the entire trait space. 481 \nMycorrhizal colonization was less strongly associated with the first axis than with the root-hair 482 \ndominated second axis, though the bivariate correlation with cortex fraction was strong as expected 483 \nin the concept of the collaboration gradient. Since mycorrhizal colonization was measured on the 484 \nsame microscopy slides as root-hair incidence, while root-hair length and cortex fraction were 485 \nmeasured separately, we do not expect a methodological bias here. As for the correlation with %M, 486 \nboth traits of the first and the second axis link to the functional concept of collaboration. This is in 487 \nline with the scheme proposed by Wen et al. (2019) who found species to either rely on a root 488 \nmorphology of high absorptive surface, which can be achieved in different ways (in their case by high 489 \nSRL or branching) or on mycorrhizal symbiosis and a high root diameter for P-scavenging. Root hairs, 490 \nas\n another absorptive structure added to the concept, are also involved in P-mining by exudation 491 \n(Wen et al., 2022) but the respective impact on their association with the collaboration gradient is 492 \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 14, 2026. ; https://doi.org/10.64898/2026.05.13.723781doi: bioRxiv preprint \n\n20 \n \nyet to be explored as P-mobilizing exudates were not measured in the current study. Taken together, 493 \nit remains to be tested on a larger set of species, if the inclusion of root hair traits widens the 494 \ncollaboration gradient to a plane encompassing different strategies for do-it-yourself resource 495 \nacquisition.  496 \nThe third axis resembled the conservation gradient proposed as the belowground analogue of the 497 \nfast-slow economic spectrum in leaves (Weigelt et al., 2021), with RTD representing the ‘slow’ and 498 \nroot-N concentration representing the ‘fast’ strategy. The degree of variation in root-hair length was 499 \nalso associated with the ‘slow’ strategy on axis three. We can only speculate that this might be 500 \nrelated to the fact that slow growing species invest in fine roots with a longer lifespan. Hence those 501 \nspecies might keep the ability to alter the length of the comparably more short-lived root hairs given 502 \nthe fact that their surface provides the main absorptive structure for those species (Fort et al., 2015). 503 \nCortex fraction loaded on the ‘fast’ side of axis three. We hypothesize that this unexpected link might 504 \noccur because AMF also enhance species N uptake under limiting conditions (Govindarajulu et al., 505 \n2005; Hodge & Fitter, 2010). As root-N concentration, cortex fraction and colonization rate were 506 \nmeasured on the same replicates, the effect of mycorrhizal colonization rate on root-N concentration 507 \nand a resulting positive correlation (r=0.3) might be overestimated by our data, which were 508 \nmeasured on plants growing under relatively low nutrient conditions. Furthermore, our experiment 509 \nwas restricted to a single AMF species and excluded plant mycorrhizal types other than arbuscular 510 \nmycorrhiza. Ectomycorrhizal species tend to occupy areas of the ‘slow’ strategy in the root 511 \neconomics space (Bergmann et al., 2020), hence adding variation to the conservation gradient that is 512 \nnot covered in the present experiment. It is also important to notice that, due to the nature of the 513 \nectomycorrhizal symbiosis with fungal hyphae covering entire fine roots and leading to fast 514 \ndegradation of root hairs (Farquhar, 1996) the importance of root-hair traits might change in a global 515 \ndataset.  516 \n 517 \nConclusions 518 \nWe can support the hypothesis that investment into root hairs and mycorrhizal partnerships are 519 \nalternative ecological strategies for soil exploration and resource uptake with a strong evolutionary 520 \nhistory. This interspecific ecological trade-off is mirrored at the intraspecific level with plants showing 521 \nmore root hairs at lower mycorrhizal colonization rates. Strong heterogeneity between species calls 522 \nfor further investigations of intraspecific patterns. A high degree of variation in root-hair incidence is 523 \na\nssociated with high mycorrhizal colonization rates and growth response at the species level. The 524 \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 14, 2026. ; https://doi.org/10.64898/2026.05.13.723781doi: bioRxiv preprint \n\n21 \n \necological trade-off between the investment in root hairs and the degree of variation in incidence 525 \nbeing intraspecifically correlated with mycorrhizal colonization rates dominates the second axis of 526 \nthe root economics space. We conclude that variation in root-hair patterns is neither fully aligned 527 \nwith the conservation gradient nor the existing concept of the collaboration gradient but rather 528 \nintroduces a new dimension of variation into the picture. Still, regarding the strong trade-off with 529 \nmycorrhizal colonization, we consider root hairs, and specifically their incidence, to add to the 530 \necological strategy of ‘do-it-yourself’. Hence, we find the concept of collaboration to span the first 531 \nand second and the conservation gradient to represent the third axis of variation in the root 532 \neconomics space. These results present strong evidence that root hairs are a considerable source of 533 \nvariation in fine root morphology that should be considered when studying belowground plant 534 \nfunctioning.  535 \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 14, 2026. ; https://doi.org/10.64898/2026.05.13.723781doi: bioRxiv preprint \n\n22 \n \nAcknowledgements 536 \nWe thank Erik Faltin, Cathrin Schierenbeck, Anja Wulf, Max Fussan, Maxi Bergmann and many others 537 \nfor help with root washing and scanning. We further thank Julien Bachelier for use of the 538 \nfluorescence microscope and in particular Maria Schauer for sharing her knowledge about fixation 539 \nand microscopy of plant material. 540 \nWe also thank the managers of the three Biodiversity Exploratories, Konstanz Wells, Swen Renner, 541 \nKirsten Reichel-Jung, Sonja Gockel, Kerstin Wiesner, Katrin Lorenzen, Andreas Hemp, Martin Gorke 542 \nand Miriam Teuscher, and all former managers for their work in maintaining the plot and project 543 \ninfrastructure; Christiane Fischer for giving support through the central office, Andreas Ostrowski for 544 \nmanaging the central data base, and Markus Fischer, Eduard Linsenmair, Dominik Hessenmöller, 545 \nDaniel Prati, Ingo Schöning, François Buscot, Ernst-Detlef Schulze, Wolfgang W. Weisser and the late 546 \nElisabeth Kalko for their role in setting up the Biodiversity Exploratories project. The work has been 547 \nfunded by the DFG Priority Program 1374 \"Infrastructure-Biodiversity-Exploratories\". Field work 548 \npermits were issued by the responsible state environmental offices of Baden-Württemberg, 549 \nThüringen, and Brandenburg. We acknowledge funding from the German Research Foundation (DFG, 550 \ngrants 432975993 to JB, KL 1866/12-1 to MvK and 323522591 to MR). 551 \n 552 \nAuthor contribution 553 \nJB designed and performed the experiment, ran the analyses and wrote the paper. TL contributed to 554 \nthe analysis and the conceptual development of the study and revised the paper. KB and EB 555 \nparticipated in the experiment and data exploration. EM revised the paper. MvK and MR contributed 556 \nto the study design and revised the paper. 557 \n 558 \nData availability 559 \nThis work is based on data elaborated by the RootFun project (323522591) and further analyzed 560 \nwithin the HAIRphae project (432975993) of the Biodiversity Exploratories program (DFG Priority 561 \nProgram 1374). All data used is publicly available in the Biodiversity Exploratories Information System 562 \n(http://doi.org/10.17616/R32P9Q) under dataset ID ##### (will be linked upon acceptance) or in 563 \ndatabases referenced. The authors explicitly encourage appropriate usage and database 564 \nimplementation of the data.  565 \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 14, 2026. ; https://doi.org/10.64898/2026.05.13.723781doi: bioRxiv preprint \n\n23 \n \nReferences 566 \nBates TR, Lynch JP. 1996. Stimulation of root hair elongation in Arabidopsis thaliana by low 567 \nphosphorus availability. Plant, Cell and Environment 19: 529–538. 568 \nBates TR, Lynch JP. 2000a. The efficiency of Arabidopsis thaliana (Brassicaceae) root hairs in 569 \nphosphorus acquisition. American Journal of Botany 87: 964–970. 570 \nBates TR, Lynch JP. 2000b. Plant growth and phosphorus accumulation of wild type and two root hair 571 \nmutants of Arabidopsis thaliana (Brassicaceae). American Journal of Botany 87: 958–963. 572 \nBergmann J, Weigelt A, Van der Plas F, Laughlin DC, Kuyper TW, Guerrero-Ramirez N, Valverde-573 \nBarrantes OJ, Bruelheide H, Freschet GT, Iversen CM, et al. 2020. The fungal collaboration gradient 574 \ndominates the root economics space in plants. Science Advances 6: 1–9. 575 \nBetekhtina AA, Tukova DE, Veselkin D V. 2023. Root structure syndromes of four families of 576 \nmonocots in the Middle Urals. Plant Diversity 45: 722–731. 577 \nBhat KKS, Nye PH. 1973. Diffusion of phosphate to plant roots in soil - I. Quantitative 578 \nautoradiography of the depletion zone. Plant and Soil 38: 161–175. 579 \nBolan NS, Robson AD, Barrow NJ. 1987. Effects of vesicular-arbuscular mycorrhiza on the availability 580 \nof iron phosphates to plants. Plant and Soil 99: 401–410. 581 \nBrown LK, George TS, Barrett GE, Hubbard SF, White PJ. 2013a. Interactions between root hair 582 \nlength and arbuscular mycorrhizal colonisation in phosphorus deficient barley (Hordeum vulgare). 583 \nPlant and Soil 372: 195–205. 584 \nBrown LK, George TS, Dupuy LX, White PJ. 2013b. A conceptual model of root hair ideotypes for 585 \nfuture agricultural environments: what combination of traits should be targeted to cope with limited 586 \nP availability? Annals of Botany 112: 317–330. 587 \nBrundrett MC. 2021. Auditing data resolves systemic errors in databases and confirms mycorrhizal 588 \ntrait consistency for most genera and families of flowering plants. Mycorrhiza: 671–683. 589 \nBrundrett MC, Tedersoo L. 2018. Evolutionary history of mycorrhizal symbioses and global host plant 590 \ndiversity. New Phytologist 220: 1108–1115. 591 \nBrundrett M, Tedersoo L. 2019. Misdiagnosis of mycorrhizas and inappropriate recycling of data can 592 \nlead to false conclusions. New Phytologist 221: 18–24. 593 \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 14, 2026. ; https://doi.org/10.64898/2026.05.13.723781doi: bioRxiv preprint \n\n24 \n \nCarminati A, Passioura JB, Zarebanadkouki M, Ahmed MA, Ryan PR, Watt M, Delhaize E. 2017. Root 594 \nhairs enable high transpiration rates in drying soils. New Phytologist 216: 771–781. 595 \nCosme M, Fernández I, van der Heijden MGA, Pieterse CMJ. 2018. Non-mycorrhizal plants: The 596 \nexceptions that prove the rule. Trends in Plant Science: 1–11. 597 \nDallstream C, Soper FM. 2024. Integrating edaphic gradients and community assembly concepts into 598 \nthe multidimensional root trait space. New Phytologist 243: 509–512. 599 \nDelhaize E, James RA, Ryan PR. 2012. Aluminium tolerance of root hairs underlies genotypic 600 \ndifferences in rhizosheath size of wheat (Triticum aestivum) grown on acid soil. New Phytologist 195: 601 \n609–619. 602 \nDíaz S, Kattge J, Cornelissen JHC, Wright IJ, Lavorel S, Dray S, Reu B, Kleyer M, Wirth C, Prentice IC, 603 \net al. 2016. The global spectrum of plant form and function. Nature 529: 167–171. 604 \nDing JX, Kong D, Zhang Z, Cai Q, Xiao J, Liu Q, Yin H. 2020. Climate and soil nutrients differentially 605 \ndrive multidimensional fine root traits in ectomycorrhizal-dominated alpine coniferous forests. 606 \nJournal of Ecology 108: 2544–2556. 607 \nDowle M, Srinivasan A. 2020. data.table: Extension of `data.frame`. https://cran.r-608 \nproject.org/package=data.table. 609 \nDurka W, Michalski SG. 2012. Daphne: a dated phylogeny of a large European flora for 610 \nphylogenetically informed ecological analyses. Ecology 93: 2297. 611 \nEissenstat DM. 1992. Costs and benefits of constructing roots of small diameter. Journal of Plant 612 \nNutrition 15: 763–782. 613 \nFarquhar L. 1996. Root hairs: specialized tubular cells extending root surfaces. The Botanical Review 614 \n62: 1–40. 615 \nFischer M, Bossdorf O, Gockel S, Hansel F, Hemp A, Hessenmoller D, Korte G, Nieschulze J, Pfeiffer 616 \nS, Prati D, et al. 2010. Implementing large-scale and long-term functional biodiversity research: The 617 \nBiodiversity Exploratories. Basic and Applied Ecology 11: 473–485. 618 \nFort F, Cruz P, Catrice O, Delbrut A, Luzarreta M, Stroia C, Jouany C. 2015. Root functional trait 619 \nsyndromes and plasticity drive the ability of grassland Fabaceae to tolerate water and phosphorus 620 \nshortage. Environmental and Experimental Botany 110: 62–72. 621 \nFreschet GT, Bellingham PJ, Lyver PO, Bonner KI, Wardle D a. 2013a. Plasticity in above- and 622 \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 14, 2026. ; https://doi.org/10.64898/2026.05.13.723781doi: bioRxiv preprint \n\n25 \n \nbelowground resource acquisition traits in response to single and multiple environmental factors in 623 \nthree tree species. Ecology and Evolution 3: 1065–1078. 624 \nFreschet GT, Cornelissen JHC, van Logtestijn RSP, Aerts R. 2010. Evidence of the ‘plant economics 625 \nspectrum’ in a subarctic flora. Journal of Ecology 98: 362–373. 626 \nFreschet GT, Cornwell WK, Wardle DA, Elumeeva TG, Liu W, Jackson BG, Onipchenko VG, 627 \nSoudzilovskaia NA, Tao J, Cornelissen JHC. 2013b. Linking litter decomposition of above- and below-628 \nground organs to plant-soil feedbacks worldwide. Journal of Ecology 101: 943–952. 629 \nFreschet GT, Pagès L, Iversen C, Comas L, Rewald B, Roumet C, Klimešová J, Zadworny M, Poorter 630 \nH, Postma J. 2021a. A starting guide to root ecology: Strengthening ecological concepts and 631 \nstandardizing root classification, sampling, processing and trait measurements. New Phytologist 232: 632 \n973–1122. 633 \nFreschet GT, Roumet C, Comas LH, Weemstra M, Bengough AG, Rewald B, Bardgett RD, De Deyn 634 \nGB, Johnson D, Klimešová J, et al. 2021b. Root traits as drivers of plant and ecosystem functioning: 635 \ncurrent understanding, pitfalls and future research needs. New Phytologist 232: 1123–1158. 636 \nGahoonia TS, Care D, Nielsen NE. 1997. Root hairs and phosphorus acquisition of wheat and barley 637 \ncultivars. Plant and Soil 191: 181–188. 638 \nGahoonia TS, Nielsen NE. 1998. Direct evidence on participation of root hairs in phosphorus (32P) 639 \nuptake from soil. Plant and Soil 198: 147–152. 640 \nGarnier S. 2018. viridis: Default Color Maps from ‘matplotlib’. https://cran.r-641 \nproject.org/package=viridis. 642 \nGilroy S, Jones DL. 2000. Through form to function: Root hair development and nutrient uptake. 643 \nTrends in Plant Science 5: 56–60. 644 \nGovindarajulu M, Pfeffer PE, Jin H, Abubaker J, Douds DD, Allen JW, Bücking H, Lammers PJ, 645 \nShachar-Hill Y. 2005. Nitrogen transfer in the arbuscular mycorrhizal symbiosis. Nature 435: 819–646 \n823. 647 \nGuerrero-Ramirez N et al. 2020. Global root traits (GRooT) database. Global Ecology and 648 \nBiogeography 30: 25–37. 649 \nGuilbeault-Mayers X, Lambers H, Laliberté E. 2024. Coordination among leaf and fine-root traits 650 \nalong a strong natural soil fertility gradient. Plant and Soil. 651 \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 14, 2026. ; https://doi.org/10.64898/2026.05.13.723781doi: bioRxiv preprint \n\n26 \n \nHaling RE, Brown LK, Bengough AG, Young IM, Hallett PD, White PJ, George TS. 2013. Root hairs 652 \nimprove root penetration, root-soil contact, and phosphorus acquisition in soils of different strength. 653 \nJournal of Experimental Botany 64: 3711–3721. 654 \nHaling RE, Yang Z, Shadwell N, Culvenor RA, Stefanski A, Ryan MH, Sandral GA, Kidd DR, Lambers 655 \nH, Simpson RJ. 2016. Root morphological traits that determine phosphorus-acquisition efficiency and 656 \ncritical external phosphorus requirement in pasture species. Functional Plant Biology 43: 815–826. 657 \nHarrell FEJ. 2020. Hmisc: Harrell Miscellaneous. https://cran.r-project.org/package=Hmisc. 658 \nvan der Heijden MGA, Martin FM, Selosse M-A, Sanders IR. 2015. Mycorrhizal ecology and 659 \nevolution: the past, the present, and the future. New Phytologist 205: 1406–1423. 660 \nHijmans RJ. 2020. raster: Geographic Data Analysis and Modeling. https://cran.r-661 \nproject.org/package=raster. 662 \nHill JO, Simpson RJ, Moore  a. D, Chapman DF. 2006. Morphology and response of roots of pasture 663 \nspecies to phosphorus and nitrogen nutrition. Plant and Soil 286: 7–19. 664 \nHodge A, Fitter AH. 2010. Substantial nitrogen acquisition by arbuscular mycorrhizal fungi from 665 \norganic material has implications for N cycling. Proceedings of the National Academy of Sciences 107: 666 \n13754–13759. 667 \nHoeksema JD, Chaudhary VB, Gehring CA, Johnson NC, Karst J, Koide RT, Pringle A, Zabinski C, 668 \nBever JD, Moore JC, et al. 2010. A meta-analysis of context-dependency in plant response to 669 \ninoculation with mycorrhizal fungi. Ecology Letters 13: 394–407. 670 \nHoldaway RJ, Richardson SJ, Dickie IA, Peltzer DA, Coomes DA. 2011. Species- and community-level 671 \npatterns in fine root traits along a 120 000-year soil chronosequence in temperate rain forest. Journal 672 \nof Ecology 99: 954–963. 673 \nHope RM. 2013. Rmisc: Ryan Miscellaneous. https://cran.r-project.org/package=Rmisc. 674 \nIversen CM, McCormack ML, Powell AS, Blackwood CB, Freschet GT, Kattge J, Roumet C, Stover DB, 675 \nSoudzilovskaia NA, Valverde-Barrantes OJ, et al. 2017. A global Fine-Root Ecology Database to 676 \naddress below-ground challenges in plant ecology. New Phytologist 215: 15–26. 677 \nJakobsen I, Chen B, Munkvold L, Lundsgaard T, Zhu Y. 2005. Contrasting phosphate acquisition of 678 \nmycorrhizal fungi with that of root hairs using the root hairless barley mutant. : 928–938. 679 \nKattge J, Bönisch G, Díaz S, Lavorel S, Prentice IC, Leadley P, Tautenhahn S, Werner GDA, Aakala T, 680 \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 14, 2026. ; https://doi.org/10.64898/2026.05.13.723781doi: bioRxiv preprint \n\n27 \n \nAbedi M, et al. 2020. TRY plant trait database – enhanced coverage and open access. Global Change 681 \nBiology 26: 119–188. 682 \nKelley A. 1950. Mycotrophy in plants. Lectures on the biology of mycorrhizae and related structures. 683 \nUSA: Chronica Botánica Co., Waltham, Mass. 684 \nKong D, Wang J, Zeng H, Liu M, Miao Y, Wu H, Kardol P. 2017. The nutrient absorption–685 \ntransportation hypothesis: optimizing structural traits in absorptive roots. New Phytologist 213: 686 \n1569–1572. 687 \nKumar A, Shahbaz M, Koirala M, Blagodatskaya E, Seidel SJ, Kuzyakov Y, Pausch J. 2019. Root trait 688 \nplasticity and plant nutrient acquisition in phosphorus limited soil. Journal of Plant Nutrition and Soil 689 \nScience 182: 945–952. 690 \nLachaise T, Bergmann J, Rillig MC, van Kleunen M. 2021. Below- and aboveground traits explain local 691 \nabundance, and regional, continental and global occurrence frequencies of grassland plants. Oikos 692 \n130: 110–120. 693 \nLambers H, Raven JA, Shaver GR, Smith SE. 2008. Plant nutrient-acquisition strategies change with 694 \nsoil age. Trends in Ecology and Evolution 23: 95–103. 695 \nLambers H, Teste FP. 2013. Interactions between arbuscular mycorrhizal and non-mycorrhizal plants: 696 \ndo non-mycorrhizal species at both extremes of nutrient availability play the same game? Plant, Cell 697 \nand Environment Cell and Environment 36: 1911–1915. 698 \nLaughlin DC, Mommer L, Sabatini FM, Bruelheide H, Kuyper TW, McCormack ML, Bergmann J, 699 \nFreschet GT, Guerrero-Ramírez NR, Iversen CM, et al. 2021. Root traits explain plant species 700 \ndistributions along climatic gradients yet challenge the nature of ecological trade-offs. Nature 701 \nEcology & Evolution 5: 1123–1134. 702 \nMa X, Li X, Ludewig U. 2021. Arbuscular mycorrhizal colonization outcompetes root hairs in maize 703 \nunder low phosphorus availability. Annals of Botany 127: 155–166. 704 \nMa X, Zarebanadkouki M, Kuzyakov Y, Blagodatskaya E, Pausch J, Razavi BS. 2018. Spatial patterns 705 \nof enzyme activities in the rhizosphere: Effects of root hairs and root radius. Soil Biology and 706 \nBiochemistry 118: 69–78. 707 \nMaherali H. 2014. Is there an association between root architecture and mycorrhizal growth 708 \nresponse? New Phytologist 204: 192–200. 709 \nMaherali H. 2017. The evolutionary ecology of roots. New Phytologist 215: 1295–1297. 710 \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 14, 2026. ; https://doi.org/10.64898/2026.05.13.723781doi: bioRxiv preprint \n\n28 \n \nMartinez Arbizu P. 2019. pairwiseAdonis: Pairwise Multilevel Comparison using Adonis. R package 711 \nversion 0.3. 712 \nMarzec M, Melzer M, Szarejko I. 2015. Root hair development in the grasses: What we already know 713 \nand what we still need to know. Plant Physiology 168: 407–414. 714 \nMatthus E, Zwetsloot M, Delory BM, Hennecke J, Andraczek K, Henning T, Mommer L, Weigelt A, 715 \nBergmann J. 2025. Revisiting the root economics space — its applications , extensions and nuances 716 \nadvance our understanding of fine-root functioning. Plant and Soil. 717 \nMcCormack ML, Iversen CM. 2019. Physical and functional constraints on viable belowground 718 \nacquisition strategies. Frontiers in Plant Science 10: 1215. 719 \nMcGonigle TP, Miller MH, Evans DG, Fairchild GL, Swan JA. 1990. A new method which gives an 720 \nobjective-measure of colonization of roots. New Phytologist 115: 495–501. 721 \nMurdoch DJ, Chow ED. 2024. ellipse: Functions for drawing ellipses and ellipse-like confidence 722 \nregions. 723 \nNestler J, Wissuwa M. 2016. Superior root hair formation confers root efficiency in some, but not all, 724 \nrice genotypes upon P deficiency. Frontiers in Plant Science 7: 1935. 725 \nOrme D, Freckleton R, Thomas D, Petzoldt T, Fritz S, Isaac N, Pearse W. 2018. caper: Comparative 726 \nAnalyses of Phylogenetics and Evolution in R. https://cran.r-project.org/package=caper. 727 \nOstonen I, Püttsepp Ü, Biel C, Alberton O, Bakker MR, Lõhmus K, Majdi H, Metcalfe D, Olsthoorn 728 \nAFM, Pronk A, et al. 2007. Specific root length as an indicator of environmental change. Plant 729 \nBiosystems 141: 426–442. 730 \nParadis E, Claude J, Strimmer K. 2004. APE: Analyses of phylogenetics and evolution in R language. 731 \nBioinformatics 20: 289–290. 732 \nParasquive V, Brisson J, Guilbeault-Mayers X, Laliberté E, Chagnon PL. 2023. Contrasted root trait 733 \nresponses between saplings of an arbuscular and an ectomycorrhizal tree species in open field 734 \ncompared to forest conditions. Journal of Ecology 111: 1700–1710. 735 \nPrice NS, Roncadori RW, Hussey RS. 1989. Cotton root growth as influenced by phosphorus nutrition 736 \nand vesicular–arbuscular mycorrhizas. New Phytologist 111: 61–66. 737 \nR Core Team. 2020. R: A language and environment for statistical computing. 738 \nReich PB. 2014. The world-wide ‘fast-slow’ plant economics spectrum: A traits manifesto. Journal of 739 \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 14, 2026. ; https://doi.org/10.64898/2026.05.13.723781doi: bioRxiv preprint \n\n29 \n \nEcology 102: 275–301. 740 \nRevell L. 2012. phytools: An R package for phylogenetic comparative biology (and other things). 741 \nMethods in Ecology and Evolution 3: 217–223. 742 \nRose L. 2017. Pitfalls in root trait calculations: How ignoring diameter heterogeneity can lead to 743 \noverestimation of functional traits. Frontiers in Plant Science 8: 898. 744 \nSchweiger PF, Robson AD, Barrow NJ. 1995. Root hair length determines beneficial effect of a 745 \nGlomus species on shoot growth of some pasture species. New Phytologist 131: 247–254. 746 \nSiqueira JO, Saggin-Júnior OJ. 2001. Dependency on arbuscular mycorrhizal fungi and responsiveness 747 \nof some Brazilian native woody species. Mycorrhiza 11: 245–255. 748 \nSmith SE, Read DJ. 2008. Mycorrhizal Symbiosis. London: Academic Press. 749 \nSoetaert K. 2014. shape: Functions for plotting graphical shapes, colors. R package version 1.4.1.: 750 \nhttp://CRAN.R-project.org/package=shape. 751 \nSoudzilovskaia NA, Vaessen S, Barcelo M, He J, Rahimlou S, Abarenkov K, Brundrett MC, Gomes 752 \nSIF, Merckx V, Tedersoo L. 2020. FungalRoot: global online database of plant mycorrhizal 753 \nassociations. New Phytologist 227: 955–966. 754 \nSun XG, Tang M. 2013. Effect of arbuscular mycorrhizal fungi inoculation on root traits and root 755 \nvolatile organic compound emissions of Sorghum bicolor. South African Journal of Botany 88: 373–756 \n379. 757 \nWarnes GR, Bolker B, Lumley T. 2020. gtools: Various R Programming Tools. https://cran.r-758 \nproject.org/package=gtools. 759 \nWeigelt A, Mommer L, Andraczek K, Iversen CM, Bergmann J, Bruelheide H, Fan Y, Freschet GT, 760 \nG\nuerrero-R\namírez NR, Kattge J, et al. 2021. An integrated framework of plant form and function: the 761\n \nbelowground perspective. New Phytologist 232: 42–59. 762 \nWen Z, Li H, Shen Q, Tang X, Xiong C, Li H, Pang J, Ryan MH, Lambers H, Shen J. 2019. Trade-offs 763 \namong root morphology, exudation and mycorrhizal symbioses for phosphorus-acquisition strategies 764 \nof 16 crop species. 765 \nWen Z, White PJ, Shen J, Lambers H. 2022. Linking root exudation to belowground economic traits 766 \nfor resource acquisition. New Phytologist 233: 1620–1635. 767 \nWickham H. 2010. ggplot2: elegant graphics for data analysis. Journal of Statistical Software 35. 768 \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 14, 2026. ; https://doi.org/10.64898/2026.05.13.723781doi: bioRxiv preprint \n\n30 \n \nWickham H, François R, Henry L, Müller K. 2020a. dplyr: A grammar of data manipulation. R package 769 \nversion 1.0.2. https://cran.r-project.org/package=dplyr. 770 \nWickham H, Hester J, Chang W. 2020b. devtools: Tools to make developing R packages easier. 771 \nhttps://cran.r-project.org/package=devtools. 772 \nWilke CO. 2024. cowplot: Streamlined plot theme and plot annotations for “ggplot2”. 773 \nWright IJ, Reich PB, Westoby M, Ackerly DD, Baruch Z, Bongers F, Cavender-Bares J, Chapin T, 774 \nCornelissen JHC, Diemer M, et al. 2004. The worldwide leaf economics spectrum. Nature 428: 821–775 \n827. 776 \nWu QS, Liu CY, Zhang DJ, Zou YN, He XH, Wu QH. 2016. Mycorrhiza alters the profile of root hairs in 777 \ntrifoliate orange. Mycorrhiza 26: 237–247. 778 \nYang Z, Culvenor RA, Haling RE, Stefanski A, Ryan MH, Sandral GA, Kidd DR, Lambers H, Simpson RJ. 779 \n2015. Variation in root traits associated with nutrient foraging among temperate pasture legumes 780 \nand grasses. Grass and Forage Science 72: 93–103. 781 \nYu R, Li X, Xiao Z, Lambers H. 2020. Phosphorus facilitation and covariation of root traits in steppe 782 \nspecies. New Phytologist 226: 1285–1298. 783 \nZhao J, Guo B, Hou Y, Yang Q, Feng Z, Zhao Y, Yang X, Fan G, Kong D. 2024. Multi-dimensionality in 784 \nplant root traits: progress and challenges. Journal of Plant Ecology 17. 785 \nZhu J, Zhang C, Lynch JP. 2010. The utility of phenotypic plasticity of root hair length for phosphorus 786 \nacquisition. Functional Plant Biology 37: 313–322. 787 \n  788 \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 14, 2026. ; https://doi.org/10.64898/2026.05.13.723781doi: bioRxiv preprint \n\n31 \n \nSupporting information 789 \nFig. S1: Microscopic image of first order root of Helictrotrichon pubescens (HUDS.) DOMORT., Poaceae, 790 \nobligate mycorrhizal.  791 \nFig. S2: Pairwise correlations of all measured traits.  792 \nFig. S3: Extended phylogenetically informed principal component analysis.  793 \nFig. S4: Variation in mycorrhizal colonization (%M) between plants of different mycorrhizal status. 794 \nTable S1: Phylogenetic signal of all measured traits. 795 \nTable S2: Phylogenetically informed principal component analysis of root hair traits and %M.  796 \nTable S3: Permanova based on pairwise dissimilarities of plant functional types and mycorrhizal 797 \ntypes within the principal component analysis displayed in Table S2. 798 \nTable S4: Extended phylogenetically informed principal component analysis. 799 \nTable S5: Permanova based on pairwise dissimilarities of plant functional types and mycorrhizal 800 \ntypes within the extended principal component analysis displayed in Table S4. 801 \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 14, 2026. ; https://doi.org/10.64898/2026.05.13.723781doi: bioRxiv preprint","source_license":"CC-BY-4.0","license_restricted":false}