Uncovering the complex relationship between plant diversity, soil properties, and mycorrhizal inoculum potential in threatened Miombo woodlands

Arbuscular mycorrhizal fungi (AMF) are key members of soil microbial communities and play a key role in regulating ecosystem processes. However, the mutual interdependence of plants and AMF in threatened Miombo woodlands has not yet been fully elucidated. In this study, we investigated the relationship between plant ecological indicators and soil mycorrhizal inoculum potential (MIP) in the Miombo forest fallow of Haut‐Katanga, Democratic Republic of Congo, and identified plant species that positively influence soil MIP. We conducted a floristic inventory on 32 plots and collected soil cores for physicochemical and AMF characterization. Trap cultures using Crotalaria juncea seedlings were performed to determine the soil MIP. We also tested the AMF colonization status of randomly collected living roots of mature woody and dominant herbaceous species to identify explanatory variables for MIP. Our results showed that MIP was twice as high in silty clay soils (63.13%) than in clay soils (30%). Furthermore, MIP increased accordingly with the relative abundance of a group of woody and herbaceous species (e.g., Albizia adianthifolia, Baphia bequaertii, and Setaria pumila) and decreased with others (e.g., Combretum collinum, Harungana madagascariensis and Hyparrhenia diplandra). Linear regressions showed that MIP increased significantly with the specific richness of woody species identified as indicators and with the amount of annual herbs. Woody legumes with high root colonization by AMF appeared to be refuge plants, and primary AMF dispersal vectors, increasing soil MIP. This study provides baseline data that can be used to formulate ecological restoration strategies, including soil and vegetation protection.


| INTRODUCTION
Arbuscular mycorrhizal fungi (AMF) are highly prevalent in terrestrial ecosystems, colonizing the roots of approximately 72% of host plant species (Brundrett & Tedersoo, 2018).The potential of a soil to support the establishment of mycorrhizal associations is defined as the mycorrhizal inoculum potential (MIP) (Brundrett & Abbott, 1995).
Mycotrophic shrubs, trees, and herbaceous species are important for maintaining the soil MIP in degraded forest patches and for improving the establishment and growth of planted seedlings (Duponnois et al., 2013).Therefore, the MIP is a fundamental feature of soil health and fertility, especially in the context of restoration of threatened forest.Indeed, AMF can improve seedling fitness by linking them to the existing mycelial network and providing them with resources (van der Heijden, 2004).Various methods, such as spore identification and counting (Abbott & Robson, 1991), serial dilution methods (Sene et al., 2012;Sene et al., 2012) and bioassay experiments in which suitable bait plants (i.e.trap cultures) are grown for a limited time in intact cores of field soil (Bois et al., 2005;Brundrett & Abbott, 1995), have been used to determine the ability of AMF propagules in soil to colonize roots.
In tropical degraded open forests, such as the Miombo woodlands, Högberg and Piearce (1986) observed that woody species are predominantly colonized by AMF.These woody species ensure the dispersal (Lies et al., 2017) and survival (Fortin et al., 2015) of AMF, thereby regulating the MIP of soils (Benkhoua et al., 2017;Hafidi et al., 2013).Previous studies have been condcted on the mycorrhizal status (AMF vs. ectomycorrhizal) (Högberg & Piearce, 1986) and diversity (Jefwa et al., 2012;Rodríguez-Echeverría et al., 2017) in tropical forest fallows.However, the interdependence between plants and AMF in threatened Miombo woodlands has not yet been fully elucidated.Therefore, identification of the biotic (indicator plant species) or abiotic (soil physicochemical parameters) factors that most influence the MIP of the Miombo soil forest is needed to improve natural regeneration or reforestation efforts, as has been done in African Mediterranean regions (Duponnois et al., 2013).
The Miombo woodlands are a vital band of tropical seasonal forest stretching from Angola, southern Democratic Republic of Congo (DRC), Malawi, parts of Zambia and Zimbabwe, to Mozambique and Tanzania (Kaumbu et al., 2021(Kaumbu et al., , 2023;;Timberlake et al., 2010).They are important habitats for a wide range of wildlife, including iconic animals such as elephants, antelopes and primates.They also provide important ecosystem services such as carbon sequestration, soil conservation and water regulation.Miombo forests are also an important source of timber, fuelwood and non-timber forest products for local communities.
They support the livelihoods of approximately 100 million people, mainly through agriculture and energy production (Kaumbu et al., 2021(Kaumbu et al., , 2023;;Kiruki et al., 2017;Syampungani et al., 2016).However, these forests are under increasing threat and pressure from deforestation and related land management issues such as erosion.There is therefore a need to restore Miombo forests due to deforestation and subsequent land management issues, such as erosion.
Miombo woodland soils are typically acidic (Bauman et al., 2016;Mapanda et al., 2013), low in available phosphorus and organic matter, and high in aluminium and iron (Bauman et al., 2016).Given these edaphic constraints, and the fact that Miombo woodlands are governed by total annual rainfall, the development and productivity of woody seedlings is expected to be highly dependent on soil microorganisms, particularly AMF (Duponnois et al., 2013).
The objective of this study was to investigate the relationship between plant ecological indicators and the MIP of the Miombo forest soils, and to identify plant species that positively influence the MIP.Specifically, we focused on woody and herbaceous species present in the savanna areas resulting from deforestation activities in two forest fallow areas, Mikembo and Kipopo located in Haut-Katanga Province in southeastern Democratic Republic of Congo.To achieve this objective, we conducted two observational studies.The first observational study involved a floristic inventory to document the plant species present in the area, while the second study investigated the interdependence between above-and below-ground communities.This included characterization of MIP, soil physicochemical properties, and AMF colonization of field roots.To the best of our knowledge, this is the first study to attempt to identify these relationships in the Miombo forest fallow.The baseline data obtained from this study will be useful in formulating strategies for ecological restoration of threatened forests, including soil conservation.the Köppen-Geiger Cwa climate classification (Peel et al., 2007).The climate is characterized by a 185-day dry season (April to October) and an average annual rainfall of 1200 mm, peaking in December.

| Description of the study area
The average annual temperature is 20 C, with the warmest months being October and November.The average annual humidity is 68%, with a monthly average of 47% during the dry season in September, and 86% during the rainy season in February (Mpundu, 2010).
Ferralsols dominate at both sites, with a high proportion of haplic Ferralsols and a low occurrence of plinthic and rhodic Ferralsols (Ngongo et al., 2009).These soils are acidic in nature with high concentrations of aluminium and iron (Bauman et al., 2016).They are also nutrient deficient with low concentrations of nitrogen and phosphorus (Bauman et al., 2016;Mpundu, 2010).Lawton (1978) classified plant species in the Miombo woodlands into four ecological groups according to their successional status: Chipya, Uapaca spp., Miombo and Mateshi.This concept of ecological groups was based on the dynamics of natural regeneration of degraded woodlands.According to Lawton (1978), the Chipya group consists of pioneer species that are heliophilic and resistant to fire and grass competition.This is the case for species in the genus Combretum (Combretaceae) and the genus Pterocarpus (Fabaceae, Papilionoïdeae).On the other hand, species of the Miombo group (Brachystegia and Julbernardia) belong to the family Fabaceae (often legumes without nodules), subfamily Caesalpinïoideae.They are shade-tolerant and require protection against fire and competition from grasses to progress from the seedling stage to the canopy (Kaumbu et al., 2021).

| Floristic inventory
The floristic inventory was conducted at the two sites, Kipopo and Mikembo, using a comprehensive design of 16 plots (4 transects) at each site.Four inventory plots were established for each transect at the end of the rainy season.To estimate tree density and herbaceous species abundance, we established inventory plots at 50 m intervals along each transect.Woody species were inventoried using a 10 m Â 10 m square quadrat, while herbaceous species were inventoried using a 1 m Â 1 m square quadrat, placed at five points in the main plot.The number of stems of each identified woody species was counted to determine the percentage relative density (RD %, number of individuals of a species divided by the number of all individuals of all species).The relative abundance-dominance cover (AD %) of herbaceous layers was determined using the quadrat method.Spatial geographic coordinates (longitude and latitude) and elevation were recorded at each site.Species were identified in the field while unidentified specimens were collected and taken to the Botany Laboratory of the Faculty of Agricultural Sciences, University of Lubumbashi, DRC, for further examination and identification.

| Soil sampling
Soil samples were collected along four parallel transects that were 150 m long and 50 m apart.Four soil cores were collected at regular intervals of 50 m along each transect using 10 cm Â 11 cm polyvinyl chloride (PVC) tubes.A total of 64 soil samples were collected per site, placed in polyethylene bags, and transported on ice to the Faculty of Agricultural Sciences, University of Lubumbashi, DR Congo.
After removal of roots and thorough homogenization, soil samples were sieved through a 2 mm mesh and divided into two subsamples.
One subsample was stored at 4 C for characterization of soil physical and chemical properties, while the other subsample was used for experimental bioassay in the nursery.

| MIP experiment
To  (1978) according to their successional status in the dynamics of natural regeneration of degraded woodlands (Kaumbu et al., 2021).Preliminary results (data not shown) showed that AMF colonization was higher in the Chipya group for P. tinctorius roots while B. spiciformis had very low (or no) ectomycorrhizal fungal colonization.Based on these results, we decided to evaluate the effect of floristic composition on MIP of soil forest fallow (Chipya).The second experiment was a bioassay, following the methodology described by Brundrett and Abbott (1995) and Bois et al. (2005).We used intact field soil cores to determine effective AMF propagules, and used Crotaliaria juncea seedlings, which are highly mycorrhizal-dependent, as a host trap culture.

| Determination of arbuscular MIP
A randomized experimental bioassay was conducted under nursery conditions to determine the MIP of soils from the Kipopo and Mikembo sites (Figure S2).Soil origin was considered as the treatment, and plots and samples as pseudoreplicates.Plants were grown in intact cores of field soil.The soil was placed in PVC tubes and ten pregerminated C. juncea seeds were transplanted into each tube.The plants were watered daily with 25 mL of distilled tap water per tube.
After 2 weeks of growth, the seedlings were carefully removed from the tubes to prevent root colonization from one seedling to another.
The entire root system of each plant was gently rinsed with tap water to remove soil, cleared with KOH for 30 min and stained using the method of Vierheilig et al. (1998).Root colonization was observed under a compound microscope at 250Â magnification.MIP values were estimated by calculating the percentage of colonized root seedlings out of a total of ten seedlings.

| Soil physicochemical characterization
Soil texture was determined from the percentage of clay (<0.002 mm), fine silt (0.002-0.02 mm), coarse silt (0.02-0.05 mm) and sand (0.05-2 mm) fractions by the improved hydrometric method (Bouyoucos, 1962).The pH was measured electrometrically in a suspension of 10 g of soil and 20 mL distilled water.Soil chemical properties including pH, extractable cations (calcium, magnesium, potassium and sodium), cation exchange capacity (CEC) and base saturation were determined using standard procedures (Bray & Kurtz, 1945).Total aluminium, iron and phosphorus were extracted by microwave digestion with a combination of acid solutions.All elements were measured by inductively coupled plasma spectrometry (ICP Agilent 5110 SVDV).Available phosphorus (Pbray2) was extracted using the Bray and Kurtz (1945) two-reagent system (0.03 N NH 4 F: 0.1 N HCl) and analyzed by flow injection analysis (FIA) using Quikchem method 12-115-01-1-A (Zellweger Analytics, Lachat Instruments Division).Total nitrogen and organic carbon were analyzed using the Kjeldahl and Walkley-Black methods, respectively.Total carbon and sulphur were measured by high temperature combustion and infrared detection in an elemental analyzer (TruMac CNS, LECO Instruments ULC).Soil organic matter was determined directly as the weight loss during combustion.All the analyses were performed at the soil laboratory of the Faculty of Forestry, Geography and Geomatics, Université Laval.

| Determination of field root AMF colonization
Fine roots were collected, gently washed under running water, and cleared in 10% KOH at 96 C for 24 h for woody root species (Onguene & Kuyper, 2001) and 1 h for herbaceous root species (Sanon, 2005).Roots were then acidified in a 1% HCl solution and stained according to Vierheilig et al. (1998) for 30 min for woody species (Onguene & Kuyper, 2001) and for 15 min for herbaceous species (Sanon, 2005).The roots were then cut into 1-2 cm pieces and placed on slides for microscopic observation at 250Â magnification.The frequency of mycorrhizal colonization (F%) was determined according to the method of Trouvelot et al. (1986): (number of colonized root sections/total number) Â 100).

| Statistical analyses
Normality of the data was assessed using using the Shapiro-Wilk test.
Homogeneity of the variance, the independence of residuals, and linearity were assessed using the Goldfeld-Quandt ( gqtest), Durbin-Watson (dwtest) and Rainbow (raintest) tests, respectively, using the R package lmtest (Zeileis & Hothorn, 2002).MIP and geophyte percentages were first log-transformed to ensure normality, homogeneity of variance, independence of residuals and linearity of the regression.
Data on AMF colonization of field roots were first transformed by arc sin √(X/100).Comparisons between sites were made using two-way analysis of variance.Means were compared using the Tukey HSD test (p < 0.05).
MIP data were pre-processed using the arc sin √(X/100) transformation followed by analysis of variance.Sites and plots were considered as fixed and random factors, respectively.Post-hoc analysis of means was performed using the Tukey HSD test.Correlations between MIP and soil physicochemical properties were determined using Pearson's correlation coefficients, calculated using the function rcorr from in the R package Hmisc v.4.1-1 library (Harrell, 2018).Generalized linear mixed models (GLMMs) were used to determine the relationship between soil MIP and the relative abundance of woody and herbaceous species, calculated using the MRT and IndVal functions from the R packages MVPARTwrap v.0.1-9.2 (Ouellette, 2011) and labdsv 1.8.0 (Roberts, 2016).Indicator plants were used to determine dissimilarities between sites.Two MLGMs were then constructed using the vegan v2.5-2 library (Oksanen et al., 2019) to determine the relationship between soil MIP and plant indicator abundance.The models were developed by grouping woody and herbaceous species, for the two sites, separated by IndVal analysis according to the percentage of clay.To ensure the accuracy of the models, collinearity between species was tested using variance inflation factors (VIF), and species with a VIF > 5 were excluded from the models.Finally, the Poisson regression was applied to both models, using the function family = Poisson (link = "log").
Species richness (S), Shannon-Wiener (H 0 ), Simpson (Ds), and Pielou (J') diversity indices and Raunkiaer's biological types were calculated as ecological indicators for the two sites (Peet & Roberts, 2013).Species richness, Shannon-Wiener, Simpson, and Pielou indices were assessed using the relative A-D values for herbaceous species and the RD values for woody species using the vegan v2.5-2 library (Oksanen et al., 2019).Presence-absence data were used to calculate the percentage of each biological type (geophytes, hemicryptophytes, phanerophytes, and therophytes).The Mann-Whitney-Wilcoxon rank test was used to calculate the inter-site dependence (Kipopo vs. Mikembo) of physicochemical parameters and ecological indicators.All data analyses were performed using R software v3.4.4 (R Development Core Team, 2018).

| Mycorrhizal soil infectivity
Arbuscules, vesicles and intra-root hyphae were observed in C. juncea (Figure 1a-d).The AMF colonization of the roots was identified as Arum-type, characterized by terminal arbuscules and intercellular hyphae (Figure 1c,d).
The bioassay analyses showed that the density of infective propagules varied greatly between sites, and this was influenced by soil physicochemical properties (Figure 2b-e) and the relative abundance of plant indicators.Soil clay content was significantly higher at Mikembo site (40.3 ± 0.5%) than at Kipopo (26.3 ± 1.7) (Figure 2b).
The Indval analysis of 25 herbaceous and 54 woody species identified 11 (44%) herbaceous and 16 (29.6%)woody species as indicators (Table S1).The soil MIP was positively associated with the

| Relationship between plant ecological indicators and soil MIP
Analysis of the diversity indices and Raunkiaer's biological types showed significant differences between the Kipopo and Mikembo sites (Figure 4).The Kipopo site had significantly higher values for the Shannon-Wiener (p = 0.01), Simpson (p = 0.01), and Pielou (p = 0.01) indices for herbaceous species than for the Mikembo forest fallow.The same trend was observed for species richness (p = 0.05) and Pielou index (p = 0.01) for woody composition species.In addition, the abundance of therophytes was significantly higher in the Kipopo site (p = 0.001) than in the Mikembo forest fallow.However, the relative abundances of hemicryptophytes (p < 0.001) and geophytes (p = 0.08) were significantly higher at the Mikembo site than at the Kipopo site.
Soil MIP increased significantly with woody species richness and the relative abundance of therophytes, while it decreased significantly with the relative abundance of hemicryptophytes and geophytes (Figure 5).

| Mycorrhizal colonization of field roots
Only intra-radical hyphae and vesicles were observed in field roots after root staining (Figure 6).AMF colonization of field roots was influenced by host plant identity and habitat site or a combination of both.Significant differences were found between species (F (5, 24) = 86.4;p < 0.0001), forest fallow sites (F (1, 24) = 117.8;p < 0.0001), and the interaction of species by site (F (5, 24) = 27.9;p < 0.0001).For example, AMF colonization of legumes such as A. polyancatha, A. adianthifolia and P. tinctorius was 22.08% ± 7.16, 20% ± 2.7 and 16.66% ± 5.8 in the Kipopo site and 11.66% ± 1.55; 10% ± 2.04 and 5% ± 1.02 in the Mikembo forest fallow (Figure 7).Legumes were more infected than other plant species, such as C. collinum (5.41% ± 1.55 and 5% ± 1.02 in Kipopo and Mikembo forest fallow, respectively).The herbaceous species showed poor mycorrhization such as Hyparrhenia diplandra (5.83% ± 1.17 in Kipopo and 5% ± 1.02 in Mikembo) and Imperata cylindrica The objective of this study was to determine the relationship between aboveground plants and soil MIP in the Miombo forest fallow, and to identify plant species that positively influence soil MIP.We found that the density of infective propagules varied greatly between sites, and that this was influenced by the relative abundance of plant indicators and soil physicochemical properties.
Our results support the mycotrophic nature of woody legume species and their potential role in maintaining high levels of soil mycorrhizal propagules in forest fallows.The woody legume species A. adianthifolia, B. bequaertii and P. tinctorius were identified as potential refuge plants and primary vectors of AMF propagules in Kipopo and Mikembo forest fallows.Previous studies from other African countries also reported that mycotrophic plant species improved soil MIP (Benkhoua et al., 2017;Duponnois et al., 2011;Hafidi et al., 2013;Sene et al., 2012;Sene et al., 2012).In addition, most of the tropical legume species previously tested in the DR Congo showed a high dependence on mycorrhizal fungi (Bulakali et al., 1999;Khasa et al., 1990Khasa et al., , 1992)), which may contribute to soil MIP.In this study, we also found that the root systems of the woody legumes were more colonized by AMF than those of the herbaceous plants, and that positive relationships were found between soil MIP and the diversity and evenness of the woody plant species in both forest fallows.This confirms other studies focusing on soil AM and ectomycorrhizal fungi in relation to vegetation gradients (Högberg & Piearce, 1986) or AMF biodiversity (Jefwa et al., 2012;Rodríguez-Echeverría et al., 2017), which also highlighted the role of plant host composition in maintaining or shifting the development of the belowground mycorrhizal population.However, compared to the study of Belay et al. (2013), who reported an AMF colonization ranging from 12 to 67% in 12 native acacia legume species in Ethiopia, the expected root AMF colonization was not achieved, as none of them showed a root colonization greater than 25%.Therefore, the contribution of woody species in our current study may also be due to their high abundance in Miombo fallow.
Indicator plants related to the contrasting mycorrhizal capacity of soils between each site were identified.In the Lubumbashi region, and particularly in the Miombo ecosystem, the dry season lasts 7 months and the dried out herbaceous cover is usually consumed by bush fires (Malaisse, 1997;Schmitz, 1971) (Kaumbu et al., 2021).
Several soil properties also showed a positive or a negative correlation with soil MIP.Analysis of soil AMF propagules using C. juncea showed that MIP was significantly higher in the silty clay soils of Kipopo than in the clay soils of Mikembo.The high clay content at the latter site may have limited root development and hence the spread of AMF propagules.The relationship between soil clay content and soil nutrient availability was previously investigated by Mujinya et al. (2013) in Lubumbashi.Their study showed a positive correlation between soil clay and aluminium and iron content.Although these elements are important sources of free oxides in the soil, they can also be toxic to AMF.Seguel et al. (2013) reported an intra-and interspecific sensitivity of AMF to aluminium, which can reduce soil MIP.
This suggests that aluminium toxicity could be a potential cause of the reduced MIP observed in the Mikembo clay soils.In addition, Al and Fe oxides and sesquioxides are primarily responsible for P immobilization in such soils.Phosphorus may simply be too low to support sufficient plant photosynthesis and to sustain high MIP.This is supported by the fact that the Kipopo site with the highest MIP also had the highest amount of P, which contradicts the negative relationship between available P and arbuscular mycorrhiza abundance reported in many studies (Sene et al., 2012;Sene et al., 2012).The results of this study support the habitat hypothesis (Zobel & Öpik, 2014), which suggests that the relationship between plants and AMF is co-dependent.
However, further studies using metagenomic approaches on a vegetation gradient or soil gradient in the Miombo forest are needed to isolate the effects of plant cover and physicochemical parameters on the observed relationship.

| CONCLUSION
In conclusion, the density of infective propagules varied greatly between sites, and this was influenced by the mycotrophic nature and The study was conducted at two forest fallow sites, Mikembo and Kipopo, located in Haut-Katanga Province in southeastern DR Congo (FigureS1).The Mikembo site, located 34.5 km northeast of Lubumbashi, is part of a forest reserve established to conserve exotic fauna and indigenous flora.It is located on the Kasenga road at coordinates 11 28'5.22"S, 27 39'35.7"E, at an altitude of 1192 m above sea level.The Kipopo site is located on a concession of the Institut National pour l' Etude et la Recherche Agronomique (INERA), located 18 km north-west of Lubumbashi at 11 34'23.58"S, 27 24'108" E, at an altitude of 1286 m above sea level.The study area falls under To investigate the influence of plant species composition on soil MIP, we examined the AMF colonization of field roots as a variable for the dominant plants.Specifically, we collected field roots from five individual woody species per site (including four legume species: Acacia polyacantha subsp.campylacantha (Hochst.ex A. Rich.)Brenan, Albizia adianthifolia (Schumach.)W. Wight, Baphia bequaertii De Wild, P. tinctorius, and one Combretaceae: Combretum collinum) and three herbaceous species (Hyparhenia diplandra, Imperata cylindrica (L.) P. Beauv.and Bidens oligoflora (Klatt) Wild).Root sampling was conducted at the end of the rainy season in May 2018.Fine roots were sampled from three individuals per woody species at four cardinal points around the stem.For each herbaceous species, the root system of the dominant species was sampled in three plots.A minimum of 80 fine roots, greater than 5 cm in length, were collected per individual and per plot.Notably, B. bequaertii and B. oligoflora were absent from the Mikembo site and were therefore excluded from the statistical analysis between sites, as well as from the interspecific variability in root AMF colonization.

( 5 .
08% ± 1.02 in Kipopo and 2% ± 0.58 in Mikembo).The legume species A. polyancatha, A. adianthifolia and P. tinctorius, had their highest root AMF colonization in the Kipopo forest fallow, while the mycorrhizal F I G U R E 3 Generalised linear mixed models (GLMM) illustrating the relationship between soil arbuscular mycorrhizal potential (MIP) and the abundance of woody and herbaceous species.N = 16 plots per site.(a).Kipopo site (MIP = 3.4-0.05AAD-0.01BBE + 0.08CAN + 0.05 DC + 0.002MM + 0.11MK + 0.1UK + 0.08IC + 0.25SP).(b).Mikembo site (MIP = 4.5-0.1AAN+ 0.03CA -0.17CCO -0.07HM -0.01IS + 0.05SL -0.14HS -0.36PA -0.15PS.For tree species: AAD: Albizia adianthifolia, AAN: a. antunesiana, BBE: Baphia bequaertii, CA: Combretum acutifolium, CCO: C. collinum, CAN: Cussonia sp., DB: Dalbergia boehmii, DC: Diplorhynchus condylocarrpon, HM: Harungana madagascariensis, IS: Hymenocardia acida, MK: Monotes katangensis, MM: Marquesia macroura, and SL: Securidaca longipendoculata.For herbaceous species: HD: Hyparhenia diplandra, IC: Imperata cylindrica, PAC: Pteridium acquilinum, PS: Panicum maximum, and SB: Setaria pumila.[Colour figure can be viewed at wileyonlinelibrary.com]F I G U R E 4 Ecological indicators of floristic composition in two sites (Kipopo and Mikembo) of the degraded open forest (Miombo).N = 16 plots per site.For herbaceous species: (a) species richness, (b) Shannon-Wiener index, (c) Simpson index and (d) Pielou index.For woody species: (e) species richness, (f) Shannon-Wiener index, (g) Simpson index, and (h) Pielou index.For the biological spectrum of Raunkiaer: (i) phanerophytes, (j) hemicryptophytes, (k) therophytes and (l) geophytes.[Colour figure can be viewed at wileyonlinelibrary.com]F I G U R E 5 Linear relationship between arbuscular MIP and the woody species richness (a), therophytes (b), hemicryptophytes (c) and geophytes (d).N = 16 plots per site.colonization of the same species was 2 to 3 times lower for in Mikembo.The legume species B. bequaertii and B. oligoflora (Asteraceae) were found only at Kipopo and showed root AMF colonization of 17.5% ± 1.02 and 5.83% ± 1.55, respectively.The data indicate the mycotrophic nature of legume species and their potential role in maintaining soil mycorrhizal propagules in forest fallow.F I G U R E 6 AMF structures showing vesicles and hyphae in root systems of tree, shrub and herbaceous species collected from the forest fallow (Miombo).(a) A. adianthifolia, (b) A. polyacantha, (c) Bidens oligoflora, (d) C. collinum, (e) H. diplandra and (f) P. tinctorius.[Colour figure can be viewed at wileyonlinelibrary.com]F I G U R E 7 Arbuscular mycorrhizal root colonisation of woody species (AP: A. polyacantha subsp.campylacantha (Hochst.ex A. Rich.)Brenan and PT: P. tinctorius), shrubs (AA: A. adianthifolia and CC: C. collinum) and herbaceous species (HD: H. diplandra, IC: I. cylindrica) in forest fallow (Miombo).The legends K and M correspond to the Kipopo and Mikembo sites respectively.N = 3 individuals per species and per site.Different letters above the box plots indicate statistically significant differences based on Tukey's HSD test (p < 0.05).[Colour figure can be viewed at wileyonlinelibrary.com] the relative abundance of plant indicators and soil physicochemical properties.The woody species A. adianthifolia, B. bequaertii and P. tinctorius and their associated herbaceous species I. cylindrica and S. pumila were identified as potential drivers of soil MIP in Kipopo and Mikembo forest fallows.The higher AMF colonization status of the formers compared to Combretaceae, geophytes and hemicryptophytes may partly explain their association with higher soil MIP.Several soil properties showed a positive correlation with soil MIP, while acidity, aluminium and iron content, clay and fine silt were negatively correlated.Our results also highlight the importance of identifying and protecting tree species with high mycorrhizal dependence to promote soil health and biogeochemical cycling, and to enhance natural or assisted regeneration processes of open forests.This study is the first to report on the interdependence of plants, soil properties and AMF in Miombo woodlands.The results of this study provide baseline data that can be used to formulate ecological restoration strategies, including soil and vegetation protection.
The indicator species A. adianthifolia was more abundant in the Kipopo site than in the Mikembo forest fallow, while Combretum acutifolium and C. collinum were more abundant in the Mikembo site.Species belonging to the genera Albizzia and Combretum have been reported as indicator spe-  Sene et al., 2012; Sene et al., 2012).Results fromZhang et al.
(Duponnois et al., 2001;)ies.The results are consistent with a previous report byOnguene and Kuyper (2001), who demonstrated interspecific variability in AMF colonization within the same rainforest site in Cameroon.In Senegal, the positive role of the legume herb Cassia obtusifolia L. in the efficient dispersal of AMF propagules has been demonstrated in both natural and planted forests(Duponnois et al., 2001;