CIAT Research Online - Accepted Manuscript Responses of earthworm communities to crop residue management after inoculation of the earthworm Lumbricus terrestris (Linnaeus, 1758) The International Center for Tropical Agriculture (CIAT) believes that open access contributes to its mission of reducing hunger and poverty, and improving human nutrition in the tropics through research aimed at increasing the eco-efficiency of agriculture. CIAT is committed to creating and sharing knowledge and information openly and globally. We do this through collaborative research as well as through the open sharing of our data, tools, and publications. Citation: Frazão, Joana; de Goede, Ron G.M.; Salánki, Tamás E.; Brussaard, Lijbert; Faber, Jack H. ; Hedde, Mickaël; Pulleman, Mirjam M. (2019). Responses of earthworm communities to crop residue management after inoculation of the earthworm Lumbricus terrestris (Linnaeus, 1758). Applied Soil Ecology, 142: 177- 188. Publisher’s DOI: https://doi.org/10.1016/j.apsoil.2019.04.022 Access through CIAT Research Online: https://hdl.handle.net/10568/101316 Terms: © 2019. CIAT has provided you with this accepted manuscript in line with CIAT’s open access policy and in accordance with the Publisher’s policy on self-archiving. This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License. You may re-use or share this manuscript as long as you acknowledge the authors by citing the version of the record listed above. You may not change this manuscript in any way or use it commercially. For more information, please contact CIAT Library at CIAT-Library@cgiar.org. 1 Responses of earthworm communities to crop residue management after inoculation of the earthworm Lumbricus terrestris (Linnaeus, 1758) Joana Frazão, Ron G. M. de Goede, Tamas E. Salánki, Lijbert Brussaard, Jack H. Faber, Mickaël Hedde, Mirjam M. Pulleman This is a "Post-Print" accepted manuscript, which has been published in "Applied Soil Ecology". This version is distributed under the Creative Commons Attribution 3.0 Netherlands License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Please cite this publication as follows: Frazão, J., de Goede, R.G.M., Salánki, T., Brussaard, L., Faber, J.H., Hedde, M., Pulleman, M.M. (2019) Responses of earthworm communities to crop residue management after inoculation of the earthworm Lumbricus terrestris (Linnaeus, 1758). Applied Soil Ecology xx xx:xx You can download the published version at: https://doi.org/10.1016/j.apsoil.2019.04.022 2 Responses of earthworm communities to crop residue management after 1 inoculation of the earthworm Lumbricus terrestris (Linnaeus, 1758) 2 3 Joana Frazão a,* , Ron G. M. de Goede a , Tamás E. Salánki a , Lijbert Brussaard a , Jack H. Faber b , 4 Mickaël Hedde c , Mirjam M. Pulleman a, d 5 a Soil Biology Group, Wageningen University & Research, P.O. Box 47, 6700 AA Wageningen, 6 The Netherlands 7 b Wageningen Environmental Research, P.O. Box 47, 6700 AA Wageningen, The Netherlands 8 c INRA, UR 251 PESSAC, F78026 Versailles CEDEX, France 9 d International Center for Tropical Agriculture (CIAT), Km 17 Recta Cali-Palmira, Apartado 10 Aéreo 6713, Zip code: 763537 Cali, Colombia 11 * Corresponding author: joana.fta.frazao@gmail.com 12 3 Abstract 13 Earthworms are important for soil functioning in arable cropping systems and earthworm species 14 differ in their response to soil tillage and crop residue management. Lumbricus terrestris 15 (Linnaeus, 1758) are rare in intensively tilled arable fields. In two parallel field trials with either 16 non-inversion (NIT) or conventional tillage (CT), we investigated the feasibility of inoculating L. 17 terrestris under different crop residue management (amounts and placement). Simultaneously, 18 we monitored the response of the existing earthworm communities to L. terrestris inoculation 19 and to crop residue treatments in terms of earthworm density, species diversity and composition, 20 ecological groups and functional diversity. L. terrestris densities were not affected by residue 21 management. We were not able to infer effects of the inoculation on the existing earthworm 22 communities since L. terrestris also colonized non-inoculated plots. In NIT and two years after 23 trial establishment, the overall native earthworm density was 1.4 and 1.6 times higher, and the 24 epigeic density 2.5 times higher, in treatments with highest residue application (S100) compared 25 to 25% (S25) or no (S0) crop residues, respectively. Residue management did not affect 26 earthworm species composition, nor the functional trait diversity and composition, except for an 27 increase of the community weighted means of bifide typhlosolis in S0 compared to S100. In CT, 28 however, crop residues did have a strong effect on species composition, ecological groups and 29 functional traits. Without crop residues (S0), epigeic density was respectively 20 and 30% lower 30 than with crop residues placed on the soil surface (S100) or incorporated (I100). Community 31 composition was clearly affected by crop residues. Trait diversity was 2.6 to 3 times larger when 32 crop residues were provided, irrespective of placement. Crop residues in CT also resulted in 33 heavier earthworms and in a shift in the community towards species with a thicker epidermis and 34 cuticle, a feather typhlosolis shape, and a higher average cocoon production rate. We conclude 35 4 that earthworm communities under conventional tillage respond more strongly to the amount of 36 crop residue than to its placement. Under non-inversion tillage, crop residue amounts affected 37 earthworm communities, but to a smaller degree than under conventional tillage. 38 39 Key-words: arable field, tillage, crop residue availability, trait-based approach, community 40 weighted mean, Rao’s quadratic entropy41 5 1. Introduction 42 Earthworms contribute crucially to soil processes, including in arable cropping systems 43 (Edwards, 2004) and have been classified into ecological groups (Bouché, 1977) to infer effects 44 on soil functioning. Endogeic species burrow horizontally in deeper soil layers and are 45 geophagous, feeding on soil organic matter. Epigeic species inhabit the topsoil without much 46 burrowing and anecic species dig deep permanent burrows with important effects on continuous 47 burrow formation and water infiltration (Keith and Robinson, 2012). Both epigeics and anecics 48 are saprophagous and feed on plant litter on the soil surface (Curry and Schmidt, 2007). 49 Earthworm communities in arable fields are dominated by endogeics (e.g., Crittenden et al., 50 2014; Frazão et al., 2017), whereas epigeics and anecics usually occur at low densities, if at all. 51 This may result in an underperformance of earthworm-mediated soil functions that are central for 52 soil quality (Andriuzzi et al., 2015; Postma-Blaauw et al., 2006). The scarcity of epigeics and 53 anecics in arable fields is thought to be the result of intensive conventional tillage (Chan, 2001): 54 direct negative effects are exposure to predation and destruction of permanent burrows of deep-55 burrowing anecics, and indirect effects are related to crop residue incorporation into the soil 56 profile. Residue incorporation is negative for epigeics and anecics (Frazão et al., 2019), but 57 positive for endogeics, by increasing the soil organic matter in the deeper layers of the soil 58 profile. Farmers are keen on having anecics inhabiting their arable soils, due to their contribution 59 to soil structure formation and water infiltration (Andriuzzi et al., 2015; Bertrand et al., 2015). 60 Previous studies have reported the effects of the anecic Lumbricus terrestris (Linnaeus, 1758) on 61 soil porosity and other soil fauna (enchytraeids, nematodes and other earthworms) seventeen 62 years after inoculation (Nuutinen et al., 2017). 63 6 Community response to disturbance has traditionally been analysed through taxonomic 64 approaches, focussing on species richness and composition (Feld et al., 2009), and in case of 65 earthworms also through broad ecological groups. However, additional information on the 66 functional ecology of communities may reflect important patterns of community assembly and 67 species coexistence (Mouchet et al., 2010), which can be better predictors of ecosystem function 68 than taxonomic indicators (Gagic et al., 2015). In this respect, Ricotta and Moretti (2011) argued 69 that community weighted means (CWM) (Garnier et al., 2004) and Rao’s quadratic entropy 70 (RaoQ) (Botta-Dukát, 2005) represent two complementary aspects of functional composition and 71 diversity of communities, i.e. the mean and the diversity of functional traits within a given 72 species assemblage, respectively. Inoculating L. terrestris in combination with improved 73 conditions conducive to its survival, as well as stimulating epigeics through the accessibility of 74 crop residues on the soil surface could be an alternative to amend functional diversity of 75 earthworm communities in arable fields. 76 In the present study, we investigated the response of earthworm communities to crop residue 77 amount and placement in the soil profile, in arable fields under different tillage practices: 78 conventional mouldboard ploughing (hereafter ―CT‖) and non-inversion tillage (hereafter 79 ―NIT‖). Our objectives were: (i) to evaluate the feasibility of inoculating L. terrestris under 80 different crop residue management in the two tillage systems; (ii) to assess how local earthworm 81 communities (density, diversity, species composition, ecological groups, and functional 82 diversity) are affected by crop residue management and inoculation of L. terrestris. In any trait-83 based approach, one of the critical aspects is trait selection. Here, we chose traits that are 84 expected to respond to food availability and position in the soil, i.e. body weight, number of 85 7 cocoons, time to maturity, reproductive strategy, typhlosolis shape, and tegument (cuticle and 86 epidermis) thickness. 87 We hypothesised that i) the inoculation of L. terrestris would be more successful where crop 88 residues were provided on the soil surface, particularly concurring with less intensive soil 89 disturbance typical of the NIT trial; ii) crop residue management and the inoculation of L. 90 terrestris would affect the earthworm community composition, with epigeics benefitting from 91 crop residue availability on the soil surface, but being subject to competition with L. terrestris 92 where inoculated; and endogeics being facilitated by the inoculation of L. terrestris; and iii) the 93 availability of crop residue on the soil surface would favour trait diversity, as well as heavier 94 earthworms with larger reproductive output, faster developmental time, with a less complex 95 typhlosolis shape and thinner tegument. 96 97 2. Materials and Methods 98 2.1 Study site 99 In the summer of 2013, two parallel field trials were installed at the PPO Westmaas research 100 farm of Wageningen University and Research, located in the southwest of The Netherlands. The 101 trials were situated in two adjacent arable fields that differed in tillage practices since 2009: CT 102 and NIT. The CT field was mouldboard ploughed annually and the NIT field was loosened 103 without soil inversion, either with a paragrubber (2009-2012 and 2014-2015) or with a spading 104 machine (2013). Previous sampling indicated that both fields lacked L. terrestris (Frazão et al., 105 unpublished results). The soil type is a Haplic Fluvisol (WRB, 2006), developed in calcareous 106 marine deposits with a sandy clay loam texture (49% sand, 24% clay) and a pH of 7.9 in the top 107 30 cm. Daily average temperature was 10.8 °C and annual precipitation was 883 mm over the 108 8 experimental period (Royal Netherlands Meteorological Institute, 2016). The crop rotation of 109 both fields was as follows: winter wheat in 2013, followed by radish (Raphanus sativus subsp. 110 oleiferus) as cover crop, sugar beet in 2014 and winter barley in 2015. Both fields received 111 similar mineral fertilization and synthetic crop protection; no animal manure was used 112 throughout the experimental period. 113 114 2.2 Experimental design 115 In August 2013, 24 plots (6x6 m) were established in the two neighbouring tillage fields, 116 arranged in a split-plot design with two factors and replicated in four blocks. Within each block, 117 the main plots were randomly assigned to the factor L. terrestris inoculation (two levels: ―+‖, 118 with inoculation and ―–‖, without inoculation), and subplots were randomly assigned to the 119 factor crop residue application (three levels that differed per trial). In the CT field the factor crop 120 residue application comprised three levels: (i) no crop residues (hereafter ―S0‖), (ii) incorporation 121 of crop residues (hereafter ―I100‖), and (iii) soil surface applied residues (hereafter ―S100‖). In the 122 NIT field, the factor crop residue application comprised the levels (i) no residues (hereafter 123 ―S0‖), (ii) 25% of crop residues placed on the soil surface (hereafter ―S25‖), and (iii) 100% of 124 crop residues placed on the soil surface (hereafter ―S100‖) (Fig. 1A). Inherent to the tillage 125 regimes, crop residue treatments under study were not exactly the same for the NIT and CT 126 systems, as it was impossible to test an incorporated crop residue treatment under non-inversion 127 tillage. 128 The crop residue amounts used in S100 (CT and NIT trials) and I100 were the same and were 129 applied annually in both trials. We kept the crop residue types as similar as possible across the 130 years, depending on availability. In 2013 a mixture of winter wheat stubble and radish 131 9 (Raphanus sativus subsp. oleiferus) was applied, as those were the crops grown in both fields. In 132 2014 a mixture of winter wheat straw and radish was applied after the removal of sugar beet 133 residues, which was the crop harvested at the time, and in 2015 only winter barley stubbles were 134 applied. Grain crop residues were applied at a rate of 4.7 t ha -1 and radish at a rate of 1.1 t ha -1 135 (DW) in the treatments S100 and I100 of both trials. 136 In October 2013, seven weeks before Fall tillage, (sub)adults of L. terrestris (Starfood, 137 Barneveld, The Netherlands) were inoculated in the ―+‖ plots of both fields at a density of 20 138 ind. m -2 . For a week prior to inoculation, the individuals were acclimatized in tempex boxes with 139 a compost substrate provided by Starfood, at 6 °C in a climate chamber. Each individual was 140 carefully checked and the ones not appearing healthy and vigorous were discarded. In each of the 141 ―+‖ plots, a 3x3 grid with 2 m spacing (Fig. 1B) was established and around each of the 142 intersects four holes were dug to 40 cm deep, and 20 individuals of L. terrestris were placed in 143 each hole. Soil pits were moistened before and after introducing earthworms, and refilled with 144 moistened soil. The order of the plots to be inoculated was a priori randomized. To prevent 145 predation by birds, flags and cannon sounds were used and upon observing mole activity, mole 146 traps were placed in the fields. 147 148 2.3 Data collection 149 2.3.1 Earthworm sampling 150 Earthworms were sampled in Spring (May) and Fall (September) 2014 and in Fall (October) 151 2015 in the CT and NIT trials. During the first two sampling events three soil monoliths of 152 30x30x20 (lxbxd) cm were collected in each plot, whereas in the last sampling event, only two 153 monoliths were taken per plot, due to logistical constraints. After digging a monolith, 2.5 l of 154 10 allyl isothiocyanate (AITC) solution (1 ml AITC dispersed in 20 ml 2-propanol added to 10 l of 155 water and mixed thoroughly) was applied to the pit, to expel deep burrowing earthworms. 156 Andriuzzi et al. (2017) have demonstrated that this is a suitable earthworm sampling method for 157 all earthworm ecological groups in arable systems. Individuals expelled by AITC were rinsed 158 and collected alive for further laboratory work. Monoliths were stored separately in plastic bags 159 for transportation and storage in the lab at 2 °C until hand-sorting. 160 2.3.2 Earthworm sample processing and body weight measurements 161 Earthworm samples were hand-sorted in the laboratory and individuals were kept alive in pots 162 with moist paper tissue at 16 ˚C for 48 h to void the guts. After voiding of the guts, live body 163 weight and developmental stage (juvenile, subadult or adult) were recorded individually for the 164 Spring 2014 samples. Specimens were then killed in 70% alcohol and identified to species 165 immediately. For the hand-sorted individuals collected in Fall 2014 and 2015, some adjustments 166 were made to reduce sample processing time. Therefore, (part of) the individuals were stored in 167 70% alcohol immediately after voiding of the guts. In those cases, the dead body weight was 168 measured after placing the specimens in water for 10 minutes, to allow body rehydration. As in 169 Spring 2014, individuals were weighed, assigned to their developmental stage and identified to 170 species. To correct for differences in the method of body weight measurement among different 171 samplings, 20 individuals sampled in Spring 2014 (live body weights ranging from 0.1 to 1.6 g) 172 were re-weighed after being stored for two years in alcohol. A linear regression (Equation 1, 173 adjusted R 2 = 0.90; p = 1.318 10-10) was computed between the rehydrated alcohol-conserved 174 body weight of 2016 (BWethanol in Equation 1) and the live body weight of 2014 (BWlive in 175 Equation 1). 176 Equation 1 177 11 The regression coefficients in Equation 1 were used as a correction factor to express all body 178 weight values per g live weight. For the purpose of this study, only (sub)adult individuals were 179 used, given that trait values for juveniles are lacking and might differ from adult trait values. 180 Adult individuals were identified using Sims and Gerard (1999) and juveniles using (Stöp-181 Bowitz, 1969), and complete individuals, as well as heads, were considered for identification. 182 Body weight was measured for intact individuals only excluding some 12% of the sampled 183 specimens. 184 185 2.4 Functional traits 186 We assessed seven functional traits (five continuous and two categorical) (Table 1) that were 187 expected to respond to resource availability: body weight in grams (measured per individual, 188 corrected for different weighing methods at different sampling occasions – see equation 1 – and 189 averaged for each species over the study duration), average number of cocoons produced per 190 year, reproductive strategy, typhlosolis shape, average time to maturity in weeks (Hedde et al., 191 2012a), and cuticle and epidermis thickness in µm (Briones and Álvarez-Otero, 2018). Body 192 weight was used as an indicator for the condition of the individuals and relates to the energetic 193 investment in growth; reproductive strategy and number of cocoons relate to the investment in 194 reproduction, thereby reflecting the potential for population recovery after disturbance; 195 typhlosolis shape relates to the nutrient uptake efficiency (Pelosi et al., 2013); time to maturity 196 reflects the investment in individual development, and often represents a trade-off with 197 reproductive investment (Stearns, 1976); finally, tegument thickness (cuticle and epidermis) 198 reflects the burrowing ability of the species (Briones and Álvarez-Otero, 2018). 199 200 12 2.5 Data analysis 201 2.5.1 Taxonomic and ecological group approaches 202 Earthworm species densities and ecological group densities (epigeic and endogeic) of 203 subsamples were averaged per plot for each sampling period and expressed as number of 204 individuals per meter square. Shannon diversity index was computed per plot, as a measure of 205 species diversity (richness and relative abundance). 206 2.5.2 Trait-based approach 207 Functional diversity was assessed by community weighted means (CWM) and Rao’s quadratic 208 entropy (RaoQ). CWM was calculated for each trait, as the mean of trait values for each species 209 in the community, weighted by the relative abundance of the species associated with that value 210 (Lavorel et al., 2008). RaoQ was calculated for the complete set of traits as the dissimilarity 211 between pairs of species within each plot, weighted by the product of the relative abundance of 212 both species (Leps et al., 2006). 213 2.5.3 Statistical analysis 214 The taxonomic, ecological group and trait data were analysed using univariate and multivariate 215 statistics. NIT and CT trials were considered separate datasets, to avoid statistical 216 pseudoreplication (Hurlbert, 1984), since the sample size of each tillage system was only one. As 217 we were interested in the effects of inoculation of L. terrestris, we excluded this species from the 218 analyses. The univariate approach consisted of mixed linear models using crop residue 219 application and inoculation treatments as fixed factors. The structure of the split-plot design was 220 incorporated in each model by nesting the crop residue application within the inoculation factor 221 in the random factors. Several response variables were modelled for each sampling season: 222 (sub)adult earthworm density, Shannon diversity index, epigeic and endogeic densities, CWM 223 13 for each trait, and RaoQ for all traits combined. If overall linear mixed models were statistically 224 significant at p< 0.05, pairwise comparisons were computed. P-value adjustments to avoid 225 inflation type I errors were considered necessary when the interaction between the fixed effects 226 was significant due to the large number of pairwise comparisons. In those cases, post-hoc 227 adjustments (Tukey HSD) were used. Overall models’ distribution and variance assumptions 228 were inspected visually, and if needed, a variance structure was used to avoid heteroscedasticity 229 (Zuur et al., 2009). 230 The multivariate approach consisted of testing the centroid ―location‖ (Anderson, 2001) and the 231 ―dispersion‖ (Anderson, 2006) of the community’s species composition. An analogy towards the 232 CWM was made with a multivariate test of the ―CWM composition‖. The centroid ―location‖ 233 analysis is a non-parametric version of a multivariate ANOVA, whereas the ―dispersion‖ 234 analysis tests the homogeneity of multivariate dispersions (Anderson, 2006). Both analyses are 235 based on dissimilarity matrices. For the species composition analysis, a Bray-Curtis dissimilarity 236 matrix was used, after square root transformation of the earthworm density data. For the CWM 237 composition analysis a Gower dissimilarity matrix was used, allowing the combination of 238 categorical and continuous variables. If the centroid location analysis was significant, a 239 nonmetric multidimensional scaling (NMDS) was plotted to visualize the results. As in the 240 univariate analysis, crop residue application and L. terrestris inoculation were used as 241 explanatory variables, and the split-plot design structure was incorporated in a permutation 242 scheme that considered our nested design. 243 We present the Fall 2015 results in the main text of this article. As most univariate and 244 multivariate tests of Spring and Fall 2014 appeared as not significant, these are presented in the 245 Supplementary materials A (Tables S1 – S9). The raw datasets of all seasons for both 246 14 experimental trials are available in the Supplementary materials B. All analyses were performed 247 with R 3.3.1 (R CoreTeam, 2014), using packages nlme 3.1–131, lsmeans 2.27-61, FD 1.0-12, 248 ade4 1.7-6 and vegan 2.4-5. 249 250 3. Results 251 3.1 Inoculation of Lumbricus terrestris 252 L. terrestris was found in both experimental trials throughout the sampling seasons (77 253 individuals in NIT vs. 46 in CT, of which 8 and 5 individuals were (sub)adults, respectively), 254 although the patterns were erratic and unrelated to the inoculation and crop residue treatments 255 (Table 2). Furthermore, besides the inoculated (sub)adult individuals, juveniles were also found 256 (Table 2), already in Spring 2014 (just seven months after inoculation). Highest average juvenile 257 density of 9.3 ind.m -2 was recorded in Fall 2014 in NIT – S25- and in CT – S100+ (Table 2), while 258 highest average densities of (sub)adults reached 2.8 ind.m -2 in NIT – S25+ and 1.9 ind.m -2 in CT 259 – I100+, also in Fall 2014. By the end of the study, in Fall 2015, no (sub)adults of L. terrestris 260 were found in the CT trial, nor in the non-inoculated plots of the NIT trial. However, irrespective 261 of the crop residue treatments, 1.4 ind.m -2 were found in the inoculated plots of the NIT trial. 262 Juveniles were found in higher densities, particularly in the NIT trial, in erratic patterns unrelated 263 to crop residue treatments. 264 3.2 Earthworm density 265 In NIT, in Fall 2015, native earthworm (sub)adult density was higher in S100 than in S25 and S0 266 (60 % and 37%, respectively, Table 3), whereas it was not affected by the inoculation of L. 267 terrestris nor by the interaction between both factors. In CT, native earthworm (sub)adult density 268 was not affected by L. terrestris inoculation, irrespective of residue application (Table 3). 269 15 3.3 Species diversity and composition 270 Besides the inoculated L. terrestris, (sub)adult individuals of six other earthworm species were 271 found in the two tillage trials: Aporrectodea caliginosa (Savigny, 1826), Allolobophora 272 chlorotica (Savigny, 1826), Aporrectodea rosea (Savigny, 1826), Eiseniella tetraedra (Savigny, 273 1826), Lumbricus castaneus (Savigny, 1826) and Lumbricus rubellus (Hoffmeister, 1843). 274 Among them, only one individual of E. tetraedra was found in each trial in Spring 2014. L. 275 castaneus was not detected during Fall 2014 (both trials), nor in Spring 2014 in the CT trial. 276 In both trials in Fall 2015, Shannon diversity index was low (≤ 1.0) and was not affected by L. 277 terrestris inoculation, irrespective of residue application (Table 3). Furthermore, in NIT, local 278 earthworm community composition was not affected by L. terrestris inoculation, irrespective of 279 residue application, whereas in CT, earthworm community composition showed differences in 280 terms of centroid location in the multivariate dimensional space, both with respect to the crop 281 residue application and to L. terrestris inoculation (Table 4, Fig. 2). The community composition 282 showed a separation between the surface-applied (S100) and the incorporated (I100) crop residue 283 treatments vs. the treatment where no crop residues (S0) were provided. The separation between 284 L. terrestris inoculation treatments was less clear (Fig. 2), concurring with the p-value of 0.042, 285 which although significant was rather high. 286 3.4 Ecological groups’ distribution 287 The NIT trial, in Fall 2015 showed a pronounced effect of surface availability of crop residues 288 on earthworm ecological groups (Table 3). Epigeics’ density was about 2.5 times higher in S100 289 than in the other treatments. Endogeics also increased significantly with crop residue availability 290 on the soil surface, although the effect was less pronounced, and the patterns were more erratic. 291 Endogeics were about 40% more abundant in S100 than in S25, but were not significantly different 292 16 from S0 (Table 3). The inoculation of L. terrestris did not affect earthworms in terms of 293 ecological groups (Table 3). 294 In the CT trial only epigeics responded to the crop residue treatments in Fall 2015 (Table 3). 295 Epigeic density in S0 treatment was 20 and 30% lower than when residues were applied on the 296 soil surface (S100) or incorporated into the soil (I100), respectively. No significant differences in 297 density of epigeics were found between S100 and I100 (Table 3). Similarly to the findings in the 298 NIT trial, the inoculation of L. terrestris did not affect earthworms in terms of ecological groups 299 (Table 3). 300 3.5 Trait composition and diversity 301 In the NIT trial, CWM of typhlosolis shape, body weight and epidermis thickness of (sub)adult 302 earthworm species were significantly affected in Fall 2015 by crop residue availability on the 303 soil surface (Table 5). In S100, the proportion of species with a bifide typhlosolis was 304 significantly smaller (-15%) compared to absence of crop residues, whereas I100 did not differ 305 from other treatments (Table 5). Neither body weight nor epidermis thickness, although 306 significant in the overall linear models, showed significant pairwise differences among any of the 307 three crop residue treatments. 308 In the CT trial in Fall 2015, the CWM body weight, number of cocoons, typhlosolis shape, and 309 cuticle and epidermis thickness were affected by the crop residue application. The distribution of 310 reproductive strategies was modified by the inoculation of L. terrestris, and the time to maturity 311 by the interaction of both factors (Table 6). CWM of (sub)adult earthworms’ body weight was 312 larger in S100 and I100 than in the S0 (16% and 9%, respectively). CWM of the number of cocoons 313 was 22% higher in S100 than in I100, which was, in turn, 40% higher than in S0. The proportion of 314 17 species with a bifide typhlosolis was 52% and 23% higher in S0 than in S100 and I100, 315 respectively. CWM of cuticle thickness was 33% larger in S100 than in I100, and in turn, it was 316 57% larger in I100 than in S0. Epidermis thickness was 4% larger in S100 and 3% larger in I100 than 317 in S0. Inoculation of L. terrestris increased biparental reproduction in the local earthworm 318 community by 6%. Finally, an interactive effect between crop residue and inoculation of L. 319 terrestris was found for the CWM of time to maturity: it was 11% higher in S0+ treatments than 320 in I100+, and between 11 to 13% higher in I100- and S0- than in S100-. 321 Multivariate analyses showed no significant patterns in CWM composition for NIT in Fall 2015, 322 but in CT, plots with crop residues (S100 and I100) were separated from plots without (S0) (Table 7, 323 Fig. 3). Although significant, trait composition as affected by the inoculation of L. terrestris 324 (Table 7) did not show such a clear separation between plots where L. terrestris had been 325 inoculated or not (Fig. 3). 326 Regarding trait diversity in Fall 2015, RaoQ was 2.6 and 3.0 times higher in S100 and I100 than 327 when no crop residues (S0) were provided in CT, while not different in NIT (Table 8). 328 329 4. Discussion 330 4.1 Attainment of L. terrestris inoculation in arable fields 331 Particularly from a farmer’s perspective, L. terrestris was successfully inoculated in both 332 experiments, since this species has established and reproduced in both fields. However the 333 success rate depended on tillage regime. The NIT trial provided better conditions for 334 establishment of this species, considering that 1.7 times as many individuals were found 335 compared to the CT trial. Additionally, more reproduction took place in the NIT trial, as 1.7 336 times more juveniles were found compared to the CT trial. Our ratio of L. terrestris individuals 337 18 collected between the CT and the NIT trials is much smaller than that of Nuutinen et al. (2011), 338 who found an average of 0.6 ind. m -2 and 4.3 ind. m -2 in conventional tillage and no-till systems, 339 respectively. However, in their study, the time span between L. terrestris inoculation and 340 sampling was 13 years. Surprisingly, in our study, L. terrestris was also found in non-inoculated 341 plots, sometimes even at higher densities than in plots that had been inoculated. We could not 342 enclose the experimental plots with physical barriers, which would have, most likely, minimized 343 the colonization of non-inoculated plots by L. terrestris. The existence of physical barriers would 344 have hampered the use of agricultural machinery, which would not be feasible under 345 conventional agricultural practices. Instead, we maximised the distances between inoculated vs. 346 non-inoculated plots (between 21 and 30 m; Fig 1A) to prevent colonization of non-inoculated 347 plots, but unfortunately this appeared not to be sufficient. Although Mather and Christensen 348 (1988) quantified the length of the surface movement of individuals of L. terrestris at 19 m in 349 one night, Eijsackers (2011) reviewed that in grazed grasslands the population’s areal expansion 350 varied between 1.5 and 4 m yr -1 , and therefore the distances between plots in our experiments 351 were expected to be sufficient to avoid the colonization of non-inoculated plots by L. terrestris. 352 However, besides active surface dispersal, passive dispersal by tractor tires (Marinissen and van 353 den Bosch, 1992) may also have promoted the occurrence of L. terrestris in non-inoculated plots. 354 In both of the two tillage systems in Spring and Fall 2014, crop residue amount or placement had 355 no effect on L. terrestris density, suggesting that L. terrestris populations were not necessarily 356 restricted by crop residue availability, in opposition to our first hypothesis. Instead of becoming 357 established where crop residues were not limiting, it is likely that L. terrestris have burrowed 358 elsewhere and initiated movement to forage (Butt et al., 2003) in the initial phase of 359 experimentation. On the other hand, by the end of the study (i.e. Fall 2015), distribution patterns 360 19 of L. terrestris, particularly juveniles, seemed to be related to crop residues application, 361 suggesting that the response of this species to crop residue availability takes time. In the NIT 362 trial, densities of juveniles were highest with full crop residue application, as well as in the CT 363 trial, provided that residues were on the soil surface. 364 Our choice of crop residue for earthworms, both the local communities and the inoculated L. 365 terrestris was pragmatic and conformed with common agricultural rotations, i.e., wheat or barley 366 followed by radish as cover crop. Although indoor experiments have shown that earthworms can 367 have good survival rates with those food sources (Al-Maliki and Scullion, 2013; Frazão et al., 368 2019; Giannopoulos et al., 2010), there is also evidence that earthworms, and in particular L. 369 terrestris, show dislike for feeding on species belonging to the Brassicaceae family (Valckx et 370 al., 2011), when subjected to food choice experiments. However, wheat and barley straw 371 applications have been shown to increase L. terrestris densities in natural populations (Stroud et 372 al., 2016), while cover cropping with radish has shown no effects on populations of this species 373 (Stroud et al., 2017). 374 375 4.2 Crop residue management and earthworm communities 376 Our results demonstrate that the local community of adult earthworms was affected by crop 377 residue availability and position, both in NIT and CT systems, although crop residue effects were 378 not similar between the tillage types. We were not able to infer effects of the inoculation on the 379 existing earthworm communities since L. terrestris colonized non-inoculated plots via active or 380 passive dispersal. 381 In CT, neither the amount nor the position of crop residues affected (sub)adult total earthworm 382 density or Shannon diversity (Table 3). However, as long as crop residues were applied, either at 383 20 the surface or incorporated at ploughing depth, epigeics’ density was 3.5 to 5 times higher than 384 in absence of residues. A similar response was found for species composition which differed 385 between plots with and without crop residues (Fig. 2). These results suggest that under 386 conventional tillage the application of crop residues, rather than the position in the soil profile, 387 plays a larger role in shaping earthworm communities. These outcomes were unexpected as we 388 hypothesised that epigeics, being known to feed on decaying litter (Bouché, 1977; Curry and 389 Schmidt, 2007), would only profit from crop residues applied on the soil surface. Furthermore, as 390 we anticipated that the most important responses in community composition due to crop residue 391 availability would be found for epigeics, we had expected that when studying species 392 composition in the multivariate space, plots without residue would be more similar to those in 393 which the crop residue was incorporated. Incorporation of crop residues under conventional 394 tillage is often claimed as a reason for the unsuitability of arable fields for epigeics (Kladivko, 395 2001). Furthermore, in a mesocosm experiment, Frazão et al. (2019) demonstrated that the 396 growth and survival of L. rubellus was reduced when crop residues (mixture of wheat straw and 397 radish) were incorporated at 30 cm soil depth. 398 In the NIT system, crop residue amount had a pronounced effect on earthworm density as well as 399 density of epigeics (Table 3), whereas species composition did not differ among the crop residue 400 treatments, which was rather surprising (Fig. 2). Crop roots that were not removed after harvest 401 may have been a food source to the earthworm populations in the no residue treatments. 402 However, this does not explain the differences in epigeic density among crop residue treatments, 403 unless the duration of our trials was not long enough to pick effects on species composition. 404 In CT, crop residue stimulated trait diversity (Table 8) and modified the community trait profiles 405 (Table 6). However, in analogy to the ecological group and community composition analyses, 406 21 the trait based approach indicated that the location of crop residue application (soil surface and 407 incorporated) was trivial, in respect to trait diversity and CWM. The observation in the CT trial 408 that trait diversity (RaoQ) was positively affected by crop residue provision suggests some 409 degree of niche differentiation in those communities. Lower competition for resources as well as 410 higher efficiency in resource utilization have been linked to higher ecosystem function (Mason et 411 al., 2005). Applying crop residues, either on the soil surface or incorporated in the profile, 412 contributed to increased earthworm body weight, and shifted the earthworm community towards 413 species with a thicker epidermis and cuticle, a feather shaped typhlosolis, and species with 414 relative high average rates for cocoon production (Table 6). Moreover, earthworm species that, 415 on average, produce more cocoons and that have a relatively thick cuticle profited even more 416 when crop residues were applied on the surface. However, those effects were always smaller in 417 magnitude than when compared to the no residue treatments (Table 6). These findings suggest 418 that crop residue availability, irrespective of position in the soil profile, promotes earthworms 419 with better burrowing abilities (i.e., larger tegument thickness, see Briones and Álvarez-Otero 420 (2018)), higher recovery from disturbance (i.e., higher reproductive output, measured as average 421 number of cocoons), higher nutrient uptake efficiency (i.e., larger proportion of species with a 422 feathered typhlosolis, see Pelosi et al. (2013)). These characteristics may contribute to a higher 423 performance of the earthworm community (i.e., larger body weight). The suggestion of higher 424 nutrient uptake efficiency by the community is surprising, as we expected that removing and not 425 applying crop residues as a food source would select for species with high nutrient uptake 426 efficiency, i.e. species with a feather shaped typhlosolis. However, typhlosolis morphology is 427 unlikely to be the only trait to determine nutrient uptake efficiency. For example Thakuria et al. 428 (2010) highlighted that earthworm species’ gut wall-associated bacterial communities shifted 429 22 according to food sources provided, although these shifts were more strongly determined by 430 habitat type and ecological group. 431 In contrast to the CT trial, in the NIT trial crop residue treatments did not affect earthworm trait 432 diversity (Table 8) nor modified the trait profiles, with the exception of typhlosolis shape (Table 433 5), where patterns were similar to those observed in the CT trial. 434 Functional responses have been amply studied in plants (e.g., D az and Cabido, 2001), while 435 little attention has been given to soil organisms. Nevertheless, earthworm functional response to 436 disturbances has been studied, in relation to tillage intensity (Pelosi et al., 2013; Pelosi et al., 437 2016), flooding of floodplains (de Lange et al., 2013; Fournier et al., 2012), and soil pollution 438 (Hedde et al., 2012b; Pérès et al., 2011). To our knowledge, this is the first study in the field 439 focussing on earthworm functional responses to crop residue availability and position. Studies 440 that have focused on the relationship between earthworm communities and crop residue 441 availability with more traditional approaches, such as community composition, ecological groups 442 or total density are also rare (but see Eriksen-Hamel et al. (2009)). The latter authors did not find, 443 however, any differences between high vs. low crop residues input in earthworm abundance or 444 biomass. Contrary to Pelosi et al. (2013) who obtained dissimilar results with different 445 approaches in studying earthworm community responses to tillage, in our study, analysis of 446 species composition, ecological groups and trait diversity and composition resulted in consistent 447 outcomes in terms of response to crop residue availability and position in NIT and CT systems. 448 Therefore, the additional value of trait-based approaches in assessing the response of earthworms 449 to crop residues management was not fully confirmed with this study. Nevertheless, since 450 functional traits represent explicit links between biology and environment, it remains useful to 451 better understand which traits are affected by crop residues, and in that respect our trait-based 452 23 approach has added value. In general, in CT, the provision of residues had an effect on several 453 facets of earthworm communities, whereas in NIT, residue quantity had small effects on 454 earthworm communities. 455 Finally, further research should focus on the hypothesis that increasing earthworm functional 456 diversity, mediated by crop residue application, enhances soil functioning. However, earthworm 457 effects might be less straightforward, as Frazão et al. (2019) found evidence of trade-offs 458 between earthworm-mediated soil porosity and formation of large water-stable macroaggregates 459 related to crop residue placement in the soil profile. 460 461 5. Conclusions 462 Our study clearly illustrates different earthworm community responses to crop residue 463 availability in arable fields under contrasting tillage regimes. The inoculation of L. terrestris was 464 successful, but the success was inconsistently related to crop residue management. In contrast, 465 the type of tillage played an important role in terms of the success of inoculations, with less 466 intensive tillage systems providing better conditions for this species than conventional 467 mouldboard ploughing. 468 The largest differences in earthworm community responses were observed between no residues 469 vs. available residues in the CT trial when using the species composition, ecological groups and 470 trait-based approaches, whereas in the NIT trial, only the use of an ecological group approach 471 enabled us to show an effect of crop residue amount on earthworms. Our results suggest that in 472 arable fields earthworms are more affected by the amount of crop residue than by its position in 473 the soil profile. 474 475 24 Acknowledgements 476 We are thankful to PPO Westmaas, in particular Marcel Tramper and Marian Vlaswinkel, who 477 allowed us to perform these trials and helped with many of the field operations. We further thank 478 students and technicians who helped in the field and in the lab, and in particular Dr. Esperanza 479 Huerta for helping with the coordination of L. terrestris handling. Dr. Angela Straathof provided 480 valuable help in editing the text. This work is part of the research programme Biodiversiteit 481 Werkt! with project number 841.11.003, financed by the Netherlands Organisation for Scientific 482 Research (NWO). 483 25 References 484 Al-Maliki, S., Scullion, J., 2013. Interactions between earthworms and residues of differing 485 quality affecting aggregate stability and microbial dynamics. Appl. Soil Ecol. 64, 56-62. 486 Anderson, M.J., 2001. A new method for non-parametric multivariate analysis of variance. 487 Austral Ecol. 26, 32-46. 488 Anderson, M.J., 2006. Distance-based tests for homogeneity of multivariate dispersions. 489 Biometrics 62, 245-253. 490 Andriuzzi, W.S., Pulleman, M.M., Cluzeau, D., Pérès, G., 2017. Comparison of two widely used 491 sampling methods in assessing earthworm community responses to agricultural intensification. 492 Appl. Soil Ecol. 119, 145-151. 493 Andriuzzi, W.S., Pulleman, M.M., Schmidt, O., Faber, J.H., Brussaard, L., 2015. Anecic 494 earthworms (Lumbricus terrestris) alleviate negative effects of extreme rainfall events on soil 495 and plants in field mesocosms. Plant Soil 397, 103-113. 496 Bertrand, M., Barot, S., Blouin, M., Whalen, J., de Oliveira, T., Roger-Estrade, J., 2015. 497 Earthworm services for cropping systems. A review. Agron. Sustain. Dev. 35, 553-567. 498 Botta-Dukát, Z., 2005. Rao's quadratic entropy as a measure of functional diversity based on 499 multiple traits. J. Veg. Sci. 16, 533-540. 500 Bouché, M.B., 1977. Strategies lombriciennes. Ecol. Bull. 25, 122-132. 501 Briones, M.J.I., Álvarez-Otero, R., 2018. Body wall thickness as a potential functional trait for 502 assigning earthworm species to ecological categories. Pedobiologia 67, 26-34. 503 Butt, K.R., Nuutinen, V., Sirén, T., 2003. Resource distribution and surface activity of adult 504 Lumbricus terrestris L. in an experimental system. Pedobiologia 47, 548-553. 505 26 Chan, K.Y., 2001. An overview of some tillage impacts on earthworm population abundance and 506 diversity — implications for functioning in soils. Soil Till. Res. 57, 179-191. 507 Crittenden, S.J., Eswaramurthy, T., de Goede, R.G.M., Brussaard, L., Pulleman, M.M., 2014. 508 Effect of tillage on earthworms over short- and medium-term in conventional and organic 509 farming. Appl. Soil Ecol. 83, 140-148. 510 Curry, J.P., Schmidt, O., 2007. The feeding ecology of earthworms – A review. Pedobiologia 50, 511 463-477. 512 de Lange, H.J., Kramer, K., Faber, J.H., 2013. Two approaches using traits to assess ecological 513 resilience: A case study on earthworm communities. Basic Appl. Ecol. 14, 64-73. 514 D az, S., Cabido, M., 2001. Vive la diff rence: plant functional diversity matters to ecosystem 515 processes. Trends Ecol. Evol. 16, 646-655. 516 Edwards, C.A., 2004. The importance of earthworms as key representatives of the soil fauna, in: 517 Edwards, C. (Ed.), Earthworm ecology. CRC Press, Boca Raton, pp. 3-11. 518 Eijsackers, H., 2011. Earthworms as colonizers of natural and cultivated soil environments. 519 Appl. Soil Ecol. 50, 1-13. 520 Eriksen-Hamel, N.S., Speratti, A.B., Whalen, J.K., Légère, A., Madramootoo, C.A., 2009. 521 Earthworm populations and growth rates related to long-term crop residue and tillage 522 management. Soil Till. Res. 104, 311-316. 523 Feld, C.K., Martins da Silva, P., Paulo Sousa, J., De Bello, F., Bugter, R., Grandin, U., Hering, 524 D., Lavorel, S., Mountford, O., Pardo, I., Pärtel, M., Römbke, J., Sandin, L., Bruce Jones, K., 525 Harrison, P., 2009. Indicators of biodiversity and ecosystem services: a synthesis across 526 ecosystems and spatial scales. Oikos 118, 1862-1871. 527 27 Fournier, B., Samaritani, E., Shrestha, J., Mitchell, E.A.D., Le Bayon, R.-C., 2012. Patterns of 528 earthworm communities and species traits in relation to the perturbation gradient of a restored 529 floodplain. Appl. Soil Ecol. 59, 87-95. 530 Frazão, J., de Goede, R.G.M., Brussaard, L., Faber, J.H., Groot, J.C.J., Pulleman, M.M., 2017. 531 Earthworm communities in arable fields and restored field margins, as related to management 532 practices and surrounding landscape diversity. Agric. Ecosyst. Environ. 248, 1-8. 533 Frazão, J., de Goede, R.G.M., Capowiez, Y., Pulleman, M.M., 2019. Soil structure formation and 534 organic matter distribution as affected by earthworm species interactions and crop residue 535 placement. Geoderma 338, 453-463. 536 Gagic, V., Bartomeus, I., Jonsson, T., Taylor, A., Winqvist, C., Fischer, C., Slade, E.M., Steffan-537 Dewenter, I., Emmerson, M., Potts, S.G., Tscharntke, T., Weisser, W., Bommarco, R., 2015. 538 Functional identity and diversity of animals predict ecosystem functioning better than species-539 based indices. Proc. Royal Soc. Lond. 282, 20142620. 540 Garnier, E., Cortez, J., Billès, G., Navas, M.-L., Roumet, C., Debussche, M., Laurent, G., 541 Blanchard, A., Aubry, D., Bellmann, A., Neill, C., Toussaint, J.-P., 2004. Plant functional 542 markers capture ecosystem properties during secondary succession. Ecology 85, 2630-2637. 543 Giannopoulos, G., Pulleman, M.M., Van Groenigen, J.W., 2010. Interactions between residue 544 placement and earthworm ecological strategy affect aggregate turnover and N2O dynamics in 545 agricultural soil. Soil Biol. Biochem. 42, 618-625. 546 Hedde, M., Pey, B., Auclerc, A., Capowiez, Y., Cluzeau, D., Cortet, J., Decaëns, T., Deharveng, 547 L., Dubs, F., Grumiaux, F., Guernion, M., Joimel, S., Laporte, M.-A., Pasquet, A., Pelosi, C., 548 Pernin, C., Ponge, J.-F., Salmon, S., Santorufo, L., Nahmani, J., 2012a. BETSI, a complete 549 28 framework for studying soil invertebrate functional traits, XVI ICSZ - International Colloquium 550 on Soil Zoology, Coimbra, Portugal. 551 Hedde, M., van Oort, F., Lamy, I., 2012b. Functional traits of soil invertebrates as indicators for 552 exposure to soil disturbance. Environ. Pollut. 164, 59-65. 553 Hurlbert, S.H., 1984. Pseudoreplication and the Design of Ecological Field Experiments. Ecol. 554 Monogr. 54, 187-211. 555 Keith, A.M., Robinson, D.A., 2012. Earthworms as natural capital: ecosystem service providers 556 in agricultural soils. Economol. J. 2, 91-99. 557 Kladivko, E.J., 2001. Tillage systems and soil ecology. Soil Till. Res. 61, 61-76. 558 Lavorel, S., Grigulis, K., McIntyre, S., Williams, N.S.G., Garden, D., Dorrough, J., Berman, S., 559 Quétier, F., Thébault, A., Bonis, A., 2008. Assessing functional diversity in the field – 560 methodology matters! Funct. Ecol. 22, 134-147. 561 Leps, J., De Bello, F., Lavorel, S., Berman, S., 2006. Quantifying and interpreting functional 562 diversity of natural communities: practical considerations matter. Preslia 78, 481-501. 563 Marinissen, J.C.Y., van den Bosch, F., 1992. Colonization of new habitats by earthworms. 564 Oecologia 91, 371-376. 565 Mason, N.W.H., Mouillot, D., Lee, W.G., Wilson, J.B., 2005. Functional richness, functional 566 evenness and functional divergence: the primary components of functional diversity. Oikos 111, 567 112-118. 568 Mather, J.G., Christensen, O., 1988. Surface movements of earthworms in agricultural land. 569 Pedobiologia 32, 399-405. 570 29 Mouchet, M.A., Villéger, S., Mason, N.W.H., Mouillot, D., 2010. Functional diversity measures: 571 an overview of their redundancy and their ability to discriminate community assembly rules. 572 Funct. Ecol. 24, 867-876. 573 Nuutinen, V., Butt, K.R., Hyväluoma, J., Ketoja, E., Mikola, J., 2017. Soil faunal and structural 574 responses to the settlement of a semi-sedentary earthworm Lumbricus terrestris in an arable clay 575 field. Soil Biol. Biochem. 115, 285-296. 576 Nuutinen, V., Butt, K.R., Jauhiainen, L., 2011. Field margins and management affect settlement 577 and spread of an introduced dew-worm (Lumbricus terrestris L.) population. Pedobiologia 54, 578 Supplement, S167-S172. 579 Pelosi, C., Pey, B., Caro, G., Cluzeau, D., Peigné, J., Bertrand, M., Hedde, M., 2016. Dynamics 580 of earthworm taxonomic and functional diversity in ploughed and no-tilled cropping systems. 581 Soil Till. Res. 156, 25-32. 582 Pelosi, C., Pey, B., Hedde, M., Caro, G., Capowiez, Y., Guernion, M., Peigné, J., Piron, D., 583 Bertrand, M., Cluzeau, D., 2013. Reducing tillage in cultivated fields increases earthworm 584 functional diversity. Appl. Soil Ecol. 83, 79-87. 585 Pérès, G., Vandenbulcke, F., Guernion, M., Hedde, M., Beguiristain, T., Douay, F., Houot, S., 586 Piron, D., Richard, A., Bispo, A., Grand, C., Galsomies, L., Cluzeau, D., 2011. Earthworm 587 indicators as tools for soil monitoring, characterization and risk assessment. An example from 588 the national Bioindicator programme (France). Pedobiologia 54, Supplement, S77-S87. 589 Postma-Blaauw, M.B., Bloem, J., Faber, J.H., Groenigen, J.W.v., Goede, R.G.M.d., Brussaard, 590 L., 2006. Earthworm species composition affects the soil bacterial community and net nitrogen 591 mineralization. Pedobiologia 50, 243-256. 592 30 Ricotta, C., Moretti, M., 2011. CWM and Rao’s quadratic diversity: a unified framework for 593 functional ecology. Oecologia 167, 181-188. 594 Royal Netherlands Meteorological Institute, Daily weather data of the Netherlands, 595 http://www.knmi.nl/nederland-nu/klimatologie/daggegevens (Accessed November 2016) 596 Sims, R.W., Gerard, B.M., 1999. Earthworms: Notes for the identification of British species. 597 Field Studies Council, Shrewsbury. 598 Stearns, S.C., 1976. Life-history tactics: a review of the ideas. Q. Rev. Biol. 51, 3-47. 599 Stöp-Bowitz, C., 1969. A contribution to our knowledge of the systematics and zoography of 600 Norwegian earthworms (Annelida Oligochaeta: Lumbricidae). Nytt Mag. Zool. 17, 169-280. 601 Stroud, J.L., Irons, D.E., Watts, C.W., Storkey, J., Morris, N.L., Stobart, R.M., Fielding, H.A., 602 Whitmore, A.P., 2017. Cover cropping with oilseed radish (Raphanus sativus) alone does not 603 enhance deep burrowing earthworm (Lumbricus terrestris) midden counts. Soil Till. Res. 165, 604 11-15. 605 Stroud, J.L., Irons, D.E., Watts, C.W., White, R.P., McGrath, S.P., Whitmore, A.P., 2016. 606 Population collapse of Lumbricus terrestris in conventional arable cultivations and response to 607 straw applications. Appl. Soil Ecol. 108, 72-75. 608 Thakuria, D., Schmidt, O., Finan, D., Egan, D., Doohan, F.M., 2010. Gut wall bacteria of 609 earthworms: a natural selection process. ISME J. 4, 357-366. 610 Valckx, J., Pina, A.C., Govers, G., Hermy, M., Muys, B., 2011. Food and habitat preferences of 611 the earthworm Lumbricus terrestris L. for cover crops. Pedobiologia 54, Supplement, S139-612 S144. 613 WRB, I.W.G., 2006. World reference base for soil resources 2006: a framework for international 614 classification, correlation and communication. FAO, Rome. 615 31 Zuur, A., Ieno, E.N., Walker, N., Saveliev, A.A., Smith, G.M., 2009. Mixed effects models and 616 extensions in ecology with R. Springer, New York. 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 32 Figure captions 637 Fig. 1. A) Scheme of the experimental design of the CT and NIT trials and list of treatments. B) 638 Details of inoculation scheme within each + plot. 639 Fig. 2. Nonmetric multidimensional scaling (NMDS) of (sub)adult earthworm communities for 640 the main factor crop residues (panels A) and C)) and main factor inoculation of L. terrestris 641 (panels B) and D)) of the non-inversion (NIT, panels A) and B), stress = 0.13) and conventional 642 tillage trials (CT, panels C) and D), stress = 0.16), in Fall 2015. Dissimilarity between species 643 composition was determined through a Bray-Curtis distance matrix and earthworm density was 644 square root transformed. Inoculated L. terrestris was excluded from dissimilarity matrices. 645 Polygons in different colours indicate different crop residues (S100: grey, S25 / I100: white, S0: 646 black) and inoculation levels (+: black, –: grey). 647 Fig. 3. Nonmetric multidimensional scaling (NMDS) of CWM for the main factor crop residues 648 (panels A) and C)) and main factor inoculation of L. terrestris (panels B) and D)) of the non-649 inversion (NIT, panels A) and B), stress = 0.08) and conventional tillage trials (CT, panels C) 650 and D), stress = 0.05), in Fall 2015. Dissimilarity between CWM composition was determined 651 through a Gower distance matrix. Inoculated L. terrestris was excluded from dissimilarity 652 matrices. Polygons in different colours indicate different crop residues (S100: grey, S25 / I100: 653 white, S0: black) and inoculation levels (+: black, –: grey). 654 655 656 657 658 33 Figures 659 660 34 661 Fig. 1 662 35 Fig. 2 663 36 Fig. 3 664 665 666 667 668 669 670 671 37 Tables 672 Table 1 Literature acquired and measured (body weight) trait values of the species sampled in 673 both trials. Earthworm species are arranged by ecological groups (first three species are 674 endogeics; and last three are epigeics). 675 Species Mean of adult body weight (g) † No. of cocoons (per year) ‡ Reproductive strategy ‡ Typhlosolis shape ‡ Time to maturity (weeks) ‡ Cuticle thickness (µm) § Epidermis thickness (µm) § A. caliginosa 0.33 27 biparental bifide 55 0.46 34.19 A. chlorotica 0.22 27 biparental bifide 36 1.60 27.39 A. rosea 0.18 35 parthenogetic bifide 55 0.67 # 32.68 # E. tetraedra 0.08 72 parthenogetic simple 13 1.74 # 27.27 # L. castaneus 0.20 65 biparental feather 24 1.74 # 27.27 # L. rubellus 0.54 106 biparental feather 37 3.21 39.42 † measured in this study 676 ‡ Hedde et al. (2012a) 677 § Briones and Álvarez-Otero (2018) 678 # Not measured in Briones and Álvarez-Otero (2018). Expert knowledge of Prof. Dr. Maria 679 Briones 680 681 38 Table 2 Mean, standard error (SE) and occurrence in number of plots (Freq) of the density of (sub)adult and juvenile individuals of L. 682 terrestris (ind. m -2 ) in the non-inversion tillage (NIT) and conventional tillage (CT) trials, for each of the sampling times (Spring 683 2014, Fall 2014 and Fall 2015). For legend of the treatments, see Figure 1. 684 NIT trial CT trial Spring 2014 Fall 2014 Fall 2015 Spring 2014 Fall 2014 Fall 2015 Mean (SE) Freq Mean (SE) Freq Mean (SE) Freq Mean (SE) Freq Mean (SE) Freq Mean (SE) Freq (Sub)adult individuals S100- 0.0 (0.0) 0 0.9 (0.9) 1 0.0 (0.0) 0 S100- 0.0 (0.0) 0 0.9 (0.9) 1 0.0 (0.0) 0 S100+ 0.0 (0.0) 0 0.09 (0.0) 0 1.4 (1.4) 1 S100+ 0.0 (0.0) 0 0.0 (0.0) 0 0.0 (0.0) 0 S25- 0.9 (0.9) 1 0.0 (0.0) 0 0.0 (0.0) 0 I100- 0.0 (0.0) 0 0.0 (0.0) 0 0.0 (0.0) 0 S25+ 0.0 (0.0) 0 2.8 (1.5) 2 1.4 (1.4) 1 I100+ 0.9 (0.9) 1 1.9 (1.3) 2 0.0 (0.0) 0 S0- 0.0 (0.0) 0 0.0 (0.0) 0 0.0 (0.0) 0 S0- 0.0 (0.0) 0 0.0 (0.0) 0 0.0 (0.0) 0 S0+ 0.0 (0.0) 0 0.0 (0.0) 0 1.4 (1.4) 1 S0+ 0.0 (0.0) 0 0.0 (0.0) 0 0.0 (0.0) 0 Juvenile individuals S100- 1.9 (1.3) 2 1.9 (1.3) 1 8.3 (2.8) 4 S100- 0.9 (0.0) 1 0.9 (0.9) 1 1.4 (1.4) 1 S100+ 4.6 (2.1) 3 8.3 (3.1) 3 6.9 (2.9) 4 S100+ 0.9 (0.0) 1 9.3 (3.0) 4 4.2 (2.0) 3 S25- 2.8 (2.8) 1 9.3 (4.7) 3 1.4 (1.4) 1 I100- 0.9 (0.0) 1 0.9 (0.9) 1 2.8 (1.8) 2 S25+ 0.0 (0.0) 0 7.4 (4.8) 3 0.0 (0.0) 0 I100+ 0.0 (0.0) 0 6.5 (3.2) 2 1.4 (1.4) 1 39 S0- 1.9 (1.9) 1 3.7 (2.9) 2 4.2 (2.0) 3 S0- 0.0 (0.0) 0 4.6 (2.5) 2 0.0 (0.0) 0 S0+ 0.0 (0.0) 0 8.3 (5.0) 3 0.0 (0.0) 0 S0+ 0.0 (0.0) 0 5.6 (2.2) 3 1.4 (1.4) 1 685 686 40 Table 3 Mean and standard error (SE) of earthworm (sub)adult density, density of epigeics and endogeics (ind. m -2 ) and Shannon 687 diversity index of the non-inversion tillage (NIT) and conventional tillage (CT) trials in Fall 2015. For legend of the treatments, see 688 Figure 1. F-statistics and associated p-value of best fitted linear mixed model of earthworm densities and Shannon diversity index. 689 Capital letters show significant pairwise differences within the main factor Crop residue application and small letters within the main 690 factor L. terrestris inoculation. 691 NIT trial CT trial Treatments (Sub)adult density Shannon diversity Epigeics † Endogeics ‡ (Sub)adult density Shannon diversity Epigeics † Endogeics ‡ S100- 109.7 (8.6) Ba 0.9 (0.1) 30.5 (3.6) Ba 79.2 (6.9) Ba 73.6 (11.4) 1.0 (0.1) 29.2 (1.4) Ba 44.4 (10.9) S100+ 97.2 (18.2) Ba 0.7 (0.1) 23.6 (8.9) Ba 73.6 (11.6) Ba 81.9 (11.4) 0.8 (0.2) 31.9 (9.2) Ba 50.0 (9.1) S25- / I100- 66.7 (22.3) Aa 0.7 (0.2) 15.3 (7.6) Aa 51.4 (17.0) Aa 75.0 (10.3) 0.6 (0.1) 13.9 (3.6) Ba 61.1 (7.5) S25+ / I100+ 62.5 (8.9) Aa 0.8 (0.0) 6.9 (1.4) Aa 55.5 (9.4) Aa 86.1 (7.3) 1.0 (0.1) 29.2 (3.5) Ba 56.9 (9.2) S0- 70.8 (15.1) Aa 0.6 (0.1) 13.9 (5.3) Aa 56.9 (11.9) ABa 70.8 (9.2) 0.7 (0.1) 6.9 (2.7) Aa 63.9 (6.6) S0+ 80.5 (16.1) Aa 0.7 (0.2) 8.3 (4.8) Aa 72.2 (15.2) ABa 56.9 (15.3) 0.3 (0.1) 5.6 (3.9) Aa 51.4 (14.1) F P F p F p F p F p F p F p F p Crop residues 9.753 0.003 1.847 0.200 18.084 0.0002 5.800 0.017 0.859 0.448 3.616 0.059 58.560 <0.0001 0.860 0.448 41 Inoculation 0.015 0.910 0.035 0.863 2.073 0.246 0.091 0.783 0.038 0.858 0.450 0.550 2.140 0.240 0.212 0.676 Crop residues x inoculation 0.445 0.651 1.456 0.272 0.039 0.962 0.703 0.515 0.690 0.520 3.620 0.059 3.058 0.085 0.422 0.665 † Epigeic species: Lumbricus castaneus, Lumbricus rubellus 692 ‡ Endogeic species Aporrectodea caliginosa, Allolobophora chlorotica, Aporrectodea rosea 693 694 42 Table 4 F and p-values from non-parametric permutational multivariate analysis of variance 695 (Location) and from multivariate homogeneity of variances (Dispersion) of (sub)adult earthworm 696 community composition for each of the main factors (crop residues and inoculation of L. 697 terrestris) and their interaction in the case of Location, of the non-inversion tillage (NIT) and 698 conventional tillage trials (CT), for Fall 2015. Inoculated L. terrestris was excluded from 699 distance matrices. Dissimilarity matrix calculated using the Bray-Curtis distance, and densities 700 were square-root transformed. 701 NIT trial CT trial Location Dispersion Location Dispersion F p F p F p F p Crop residues 1.474 0.082 0.490 0.520 3.555 0.013 1.126 0.217 Inoculation 1.064 0.559 0.141 0.778 1.886 0.042 2.315 0.223 Crop residues x inoculation 0.335 0.794 - - 2.095 0.072 - - 702 703 43 Table 5 Mean and standard error (SE) of community weighted means (CWM) for the trait values in the non-inversion tillage trial 704 (NIT), for Fall 2015. Earthworm community taken into account for the computation excluded inoculated L. terrestris. For legend of 705 the treatments, see Figure 1. F-statistics and associated p-value of best fitted linear mixed model of CWM. Both categorical traits only 706 had two trait values, therefore, only one is shown. Capital letters show significant pairwise differences within the main factor Crop 707 residue application and small letters within the main factor L. terrestris inoculation. 708 Treatments Body weight (g) No. of cocoons (per year) Reproductive strategy † Typhlosolis shape ‡ Time to maturity (weeks) Cuticle thickness (µm) Epidermis thickness (µm) S100- 0.37 (0.01) Aa 49.02 (1.55) 0.93 (0.05) 0.72 (0.03) Aa 48.40 (1.55) 1.30 (0.11) 34.87 (0.39) Aa S100+ 0.35 (0.02) Aa 43.00 (4.77) 0.95 (0.03) 0.78 (0.06) Aa 50.37 (1.43) 1.01 (0.17) 34.70 (0.64) Aa S25- 0.36 (0.02) Aa 42.21 (6.46) 0.95 (0.04) 0.81 (0.08) ABa 50.77 (1.69) 1.04 (0.23) 34.78 (0.54) Aa S25+ 0.33 (0.01) Aa 36.99 (2.48) 0.92 (0.05) 0.88 (0.03) ABa 50.14 (1.35) 0.97 (0.06) 33.71 (0.68) Aa S0- 0.36 (0.02) Aa 41.51 (6.05) 0.96 (0.02) 0.82 (0.07) Ba 51.09 (0.90) 1.00 (0.18) 34.82 (0.58) Aa S0+ 0.33 (0.01) Aa 35.37 (4.59) 0.91 (0.04) 0.90 (0.05) Ba 51.22 (1.20) 0.87 (0.13) 33.83 (0.67) Aa F p F p F p F p F p F p F p Crop residues 4.310 0.039 3.746 0.055 0.044 0.957 4.710 0.031 1.444 0.274 1.267 0.317 4.915 0.028 Inoculation 1.860 0.266 1.239 0.347 0.801 0.437 1.217 0.351 0.103 0.770 1.321 0.334 0.902 0.412 44 Crop residues x inoculation 0.553 0.589 0.009 0.991 0.571 0.580 0.035 0.966 0.806 0.469 0.314 0.736 1.638 0.235 † Results presented for the category of biparental reproductive strategy; 709 ‡ Results presented for the category of bifide typhlosolis. 710 45 Table 6 Means and standard errors of community weighted means (CWM) for the trait in the conventional tillage trial (CT), for Fall 711 2015. Earthworm community taken into account for the computation excluded inoculated L. terrestris. For legend of the treatments, 712 see Figure 1. F-statistics and associated p-value of best fitted linear mixed model of CWM. Both categorical traits only had two trait 713 values, therefore, only one is shown. Capital letters show significant pairwise differences within the main factor Crop residue 714 application and small letters within the main factor L. terrestris inoculation. When only small letters are provided, significant 715 differences refer to the interaction between both treatments. 716 Treatments Body weight (g) No. of cocoons (per year) Reproductive strategy † Typhlosolis shape ‡ Time to maturity (weeks) Cuticle thickness (µm) Epidermis thickness (µm) S100- 0.40 (0.02) Ba 61.67 (5.21) Ca 0.89 (0.05) Aa 0.57 (0.07) Aa 46.51 (1.46) ab 1.71 (0.19) Ca 35.98 (0.53) Ba S100+ 0.39 (0.01) Ba 55.87 (7.29) Ca 0.97 (0.03) Ab 0.63 (0.10) Aa 47.26 (2.32) abcd 1.50 (0.27) Ca 35.57 (0.34) Ba I100- 0.36 (0.01) Ba 41.51 (2.27) Ba 0.93 (0.05) Aa 0.82 (0.03) Aa 51.81 (0.53) cd 0.96 (0.08) Ba 35.02 (0.19) Ba I100+ 0.38 (0.01) Ba 54.72 (4.19) Ba 0.90 (0.06) Ab 0.65 (0.06) Aa 47.65 (1.18) ac 1.47 (0.14) Ba 35.37 (0.40) Ba S0- 0.33 (0.01) Aa 34.68 (2.38) Aa 0.90 (0.01) Aa 0.91 (0.03) Ba 52.36 (1.04) cd 0.78 (0.11) Aa 34.12 (0.18) Aa S0+ 0.35 (0.01) Aa 34.24 (4.21) Aa 1.00 (0.00) Ab 0.91 (0.05) Ba 52.72 (0.79) bd 0.75 (0.13) Aa 34.44 (0.45) Aa F p F p F p F p F p F p F p Crop residues 17.000 0.0003 25.566 <0.00001 0.579 0.575 53.564 <0.0001 16.291 0.0004 13.743 0.001 19.060 0.0002 46 Inoculation 1.796 0.273 0.424 0.562 64.751 0.004 7.008 0.077 12.328 0.039 0.475 0.540 0.350 0.598 Crop residues x inoculation 0.415 0.670 2.523 0.122 1.466 0.269 2.907 0.093 4.322 0.039 2.686 0.109 0.640 0.544 † Results presented for the category of biparental reproductive strategy; 717 ‡ Results presented for the category of bifide typhlosolis. 718 47 Table 7 F and p-values from non-parametric permutational multivariate analysis of variance 719 (Location) and from multivariate homogeneity of variances (Dispersion) of CWM for each of the 720 main factors (crop residues and inoculation) and their interaction in the case of Location, of the 721 non-inversion (NIT) and conventional tillage (CT) trials, for Fall 2015. Inoculated L. terrestris 722 was excluded from distance matrices. Dissimilarity matrix calculated using the Gower distance. 723 NIT trial CT trial Location Dispersion Location Dispersion F P F p F p F p Crop residues 0.939 0.262 0.0495 0.960 9.690 0.002 1.0216 0.177 Inoculation 1.834 0.336 0.0433 0.868 1.306 0.043 0.0513 0.834 Crop residues x inoculation 0.085 0.949 - - 1.779 0.260 - - 724 48 Table 8 Mean and standard error of RaoQ in the non-inversion tillage (NIT) and conventional 725 tillage (CT) trials, for Fall 2015. Earthworm community taken into account for the 726 computation excluded inoculated L. terrestris. For legend of the treatments, see Figure 1. F-727 statistics and associated p-value of best fitted linear mixed model of RaoQ. Capital letters 728 show significant pairwise differences within the main factor Crop residue application and 729 small letters within the main factor L. terrestris inoculation. 730 Treatments NIT trial CT trial S100- 0.10 (0.01) 0.12 (0.01) Ba S100+ 0.07 (0.01) 0.10 (0.02) Ba S25- / I100- 0.06 (0.02) 0.07 (0.01) Ba S25+ / I100+ 0.06 (0.01) 0.11 (0.01) Ba S0- 0.06 (0.02) 0.04 (0.01) Aa S0+ 0.05 (0.01) 0.03 (0.02) Aa F p F p Crop residues 3.731 0.055 17.717 0.0003 Inoculation 2.756 0.196 0.138 0.735 Crop residues x inoculation 0.511 0.613 2.792 0.101 731 732 733 734 735