Agriculture, Ecosystems and Environment 134 (2009) 257–268

Distinct sporulation dynamics of arbuscular mycorrhizal fungal communitiesfrom different agroecosystems in long-term microcosms

Fritz Oehl a,b,*, Ewald Sieverding c, Kurt Ineichen b, Paul Mader d, Andres Wiemken b, Thomas Boller b

a Ecological Farming Systems, Agroscope Reckenholz-Tanikon Research Station ART, Reckenholzstrasse 191, CH-8046 Zurich, Switzerlandb Plant Science Center Zurich-Basel, Institute of Botany, University of Basel, Hebelstrasse 1, CH-4056 Basel, Switzerlandc Institute of Plant Production and Agroecology in the Tropics and Subtropics, University of Stuttgart Hohenheim, Garbenstr. 13, D-70593 Stuttgart-Hohenheim, Germanyd Research Institute of Organic Agriculture, Ackerstrasse, CH-5070 Frick, Switzerland

A R T I C L E I N F O

Article history:

Received 7 April 2009

Received in revised form 23 July 2009

Accepted 29 July 2009

Available online 26 August 2009

Keywords:

Arbuscular mycorrhiza

Community structure

Biodiversity

Life cycle

Life history strategies

Organic agriculture

Soil quality

Soil function

A B S T R A C T

The aim of this study was to investigate sporulation dynamics of arbuscular mycorrhizal fungal (AMF)

communities from agroecosystems differing in land use intensity in long-term experimental

microcosms. These were set up with characteristic grassland plants (Lolium perenne, Trifolium pratense,

Plantago lanceolata), and inoculated with soils from several grasslands and arable lands subjected to crop

rotation or continuous monocropping. The microcosms were maintained under ambient light and

temperature conditions over 3 years. A novel, localized sampling scheme was applied for attaining

exclusively the newly formed spores at bimonthly intervals. Overall, 39 AMF species were detected by

morphological spore identification. Some species were recovered from all sites, others exclusively from

arable lands, or grasslands, or from all sites except under maize monocropping. Clear seasonal and

successional AMF sporulation dynamics were revealed, implying different life strategies of different AMF

species. A first group of Glomus spp., including G. mosseae, sporulated rapidly during the first season. A

second group, including G. constrictum and G. fasciculatum, sporulated late in the first season and

replaced the first group during subsequent seasons. A large third group, including G. invermaium, G.

macrocarpum and G. sinuosum, sporulated much later, in the second or third season. Acaulospora,

Archaeospora and Ambispora spp. sporulated mainly during spring and early summer, Scutellospora and

Cetraspora spp. only in fall. While in the microcosms derived from arable lands, cumulative species

numbers did not increase anymore after 2 years, the numbers still increased significantly in the

microcosms from the grasslands indicating longer lasting periods of sporulation cycles. Remarkably, the

arable land under organic farming produced the highest AMF species richness, even higher than the

grasslands. In conclusion, AMF communities from distinct agro-ecosystems differed in species

composition and seasonal and successional sporulation dynamics.

� 2009 Elsevier B.V. All rights reserved.

Contents lists available at ScienceDirect

Agriculture, Ecosystems and Environment

journal homepage: www.e lsev ier .com/ locate /agee

1. Introduction

Arbuscular mycorrhizal (AM) fungi form a mutually beneficialsymbiosis with most herbaceous plants (Smith and Read, 2008).Experimental work has demonstrated that a diverse AMF commu-nity is crucial for the functioning of terrestrial ecosystems anddecisive for plant community structure and ecosystem productivity(van der Heijden et al., 1998a,b; Klironomos et al., 2000; Bever et al.,2001; Burrows and Pfleger, 2002; Hart et al., 2003). There is greatinterest to assess the diversity of AM fungi in both naturalecosystems and in agroecosystems with different management

* Corresponding author at: Agroscope Reckenholz-Tanikon Research Station ART,

Ecological Farming Systems, Reckenholzstrasse 191, CH-8046 Zurich, Switzerland.

Tel.: +41 44 377 7321; fax: +41 44 377 7201.

E-mail address: Fritz.Oehl@art.admin.ch (F. Oehl).

0167-8809/$ – see front matter � 2009 Elsevier B.V. All rights reserved.

doi:10.1016/j.agee.2009.07.008

practices. Most current studies addressing AMF diversity rely onmorphological identification of AMF spores extracted directly fromthe field (Douds and Millner, 1999; Landis et al., 2004; Gai et al.,2006) or, additionally, from so-called trap cultures in which soilsamples from the field sites are used as inocula to propagate theindigenous AMF populations on the roots of appropriate host plantsin pot cultures (Jansa et al., 2002; Oehl et al., 2003a, 2005b;Mathimaran et al., 2005). However, these techniques have theirconstraints as spore density assessed in field samples not necessarilyreflects the AMF population currently colonizing the roots (Clappet al., 1995) and, moreover, it is often not feasible to readilydistinguish freshly formed spores from spores formed during anearlier season (Lee and Koske, 1994). Also, some AMF species mightsporulate only infrequently or not at all in the field as well as undertrap culture conditions (Clapp et al., 1995; Oehl et al., 2004).

An alternative to studying spores from soil would be to examineAM fungi within living plant roots; this would ascertain that the

F. Oehl et al. / Agriculture, Ecosystems and Environment 134 (2009) 257–268258

momentarily active part of the AMF community is sampled.However, there are no morphological criteria to identify AMFspecies within roots, with few exceptions where Glomalean genera(or families) can be distinguished (e.g. Merrywheather and Fitter,1998a,b). In principle, some of the recently developed methodsusing molecular tools to identify AMF species would be ideal tostudy AMF communities in root samples (Helgason et al., 1998;Redecker, 2000; Rodriguez et al., 2004; Stukenbrock and Rosen-dahl, 2005; Mummey et al., 2005; Opik et al., 2006). However, costsand labor of these techniques have so far prevented their large-scale application needed for a realistic assessment of AMFdiversity: for each study site, a number of root fragments shouldbe analyzed from several individuals of every plant speciesoccurring at the site, e.g. a species rich grassland, in order toobtain some confidence that the majority of the AMF speciesoccurring at the site have been detected (Redecker, 2000; Borstleret al., 2006; Hijri et al., 2006; Renker et al., 2006; Santos et al.,2006). Even then it can be argued that the results reflect only thestatus at the date of sampling, necessitating repeated analyses;moreover, they may reveal only the AMF species dominantlycolonizing the roots, which may not be identical to those mostactively contributing to symbiotic performance via the extra-radical hyphal network.

One aspect of AMF diversity that is little understood is their life-history strategies and successional dynamics (Hart et al., 2001;Pietrowski and Rillig, 2008). Initially, temporal development of AMfungi in ecosystems was studied by monitoring spore density and/or root colonization without regard for species diversity (e.g.Hayman, 1970; Giovannetti, 1985; Sylvia, 1986; Bentivenga andHetrick, 1992; Gavito and Varela, 1993; Mandyam and Jumpponen,2008). More recent work assessed spores of different AMF speciesseparately (An et al., 1993; Sturmer and Bellei, 1994; Hendrix et al.,1995) and recorded their diversity and abundance in differentsuccessional stages of grasslands (Johnson et al., 1992), sand dunes(Błaszkowski, 1994; Błaszkowski and Tadych, 2000) or duringprimary succession of volcanic deserts (Titus et al., 2007; Wu et al.,2007), riparian ecosystems (Beauchamp et al., 2006; Pietrowskiet al., 2008), or in the course of the development of a specific plant(Tetragastris panamensis) in a tropical rain forest (Husband et al.,2002). A better understanding of AMF life history strategies wouldbe of particular interest with regard to application of AM fungi forreclamation and revegetation of degraded land and for developingmore sustainable agricultural land-use systems (Miller et al., 1985;Smith et al., 1998; Cuenca et al., 1998; Richter and Stutz, 2002; Wuet al., 2002; Pietrowski and Rillig, 2008).

Asexual AMF chlamydospores, recently called glomerospores(Goto and Maia, 2006), are transitorily dormant, persistentpropagules that remain infectious in the absence of host plantsand can survive under unfavorable conditions (Klironomos andHart, 2002). Thus, spore formation may represent a crucial life-history strategy of AM fungi for surviving in periodically disturbedhabitats such as cultivated agro-ecosystems. We expected, there-fore, that AMF communities inhabiting differently managed agro-ecosystems exhibit temporal sporulation patterns reflecting theagricultural disturbances imposed by the farming practices. To testthis hypothesis, we used a consortium of typical grassland plants inlong-term microcosms for observing temporal sporulation pat-terns of AMF communities derived from differently managed agro-ecosystems. Earlier we described the emergence of spores ofdifferent AMF species within the first 8 months after set up of thesemicrocosms (Oehl et al., 2003a). Here we present the results of thesporulation pattern observed over a period of 3 years under naturalambient light and temperature conditions as achieved in agreenhouse. The agroecosystems examined included three no-to low-input, plant species rich grasslands, two cultivated sitessubjected to low- to moderate input farming with a 7-year crop

rotation and three sites where high-input continuous maizemonocropping was practiced. Soil cores were taken four timesduring the vegetative period at bimonthly intervals, each time atthe same, defined location in the microcosms whereafter thesampling holes were refilled with sterile new substrate. In thisway, only spores, which had been formed within the preceding 2months were sampled. We argued that this sampling strategyshould allow us to assess successional and seasonal sporulationdynamics of the AMF communities. We expected to recognizeseveral ‘‘ecological AMF groups’’ that differ in life strategies,wondering if different agricultural practices affect such ecologicalAMF groups differently.

2. Materials and methods

2.1. Origin of the AMF communities, field sites

The AMF communities were derived from eight field siteslocated in the plain of the Upper Rhine valley between themountainous ranges of the Black Forest, the Vosges and the Jura(located in France, Germany and Switzerland; see Oehl et al.,2003a). The AMF community composition as determined by sporemorphotyping in the field samples and in 8-months old trapcultures as well as the regional climate, soil types and some generalsoil characteristics have been described previously (Oehl et al.,2003a). According to FAO (2006) the soil types were HaplicRegosols developed on Loess at sites V and R, Haplic Luvisolsdeveloped on Loess at G, O, L, S and F, and a Rendzic Leptosoldeveloped on Jurassic limestone at site W (Oehl et al., 2003a).

The agricultural use of the sites ranges from semi-natural,extensive grasslands (site W, V and G), to arable lands with a 7-yearcrop rotation managed organically and conventionally at low- tomoderate farming intensity (sites O and L, where managementcomplied with the rules of Swiss bio-organic farming and SwissIntegrated Production, respectively), to lands with intensivelymanaged, high-input maize monocropping (sites S, F and R). Theagricultural practices applied have been described in detail (Oehlet al., 2003a). Site O is part of the Swiss DOC long-term fieldexperiment, in which organic and conventional farming practiceshave been compared since 1978 (Mader et al., 2002; Oehl et al.,2004).

2.2. AMF inocula from the field sites

At each of the eight field sites, four replicate plots (plot size5 m � 20 m) were chosen according to the plot sizes given at site O(Mader et al., 2002, see above). The field replicates thusrepresented pseudo-replicates at 7 of 8 sites. From each of thereplicate plots, six soil cores (10 cm depth) were taken in March2000 (Oehl et al., 2003a). The AMF inoculum used to set up themicrocosms consisted of undisturbed soil crumbs from 5 to 7 cmsoil depth (i.e. intact soil aggregates of 3–8 cm3 volume; in total180 g per microcosm) obtained from the six soil cores collected perreplicate plot.

2.3. Microcosms for AM fungi

With each AMF inoculum, representing one of the four replicateplots at each field site, two experimental ‘microcosms’(300 mm� 200 mm� 200 mm soil volume) were set up in thegreenhouse. The microcosms represented a model grasslandconsisting of a grass, legume and forb consortium (Lolium perenne,

Trifolium pratensis, Plantago lanceolata). Eight non-mycorrhizalmicrocosms with the same plant consortium plus a supplementof autoclaved soil inoculum from site G were also included, yieldinga total of 72 microcosms. The set-up and management of these

F. Oehl et al. / Agriculture, Ecosystems and Environment 134 (2009) 257–268 259

microcosms (AMF trap cultures) has been described in detailpreviously (Oehl et al., 2003a). Briefly, the same autoclaved mixtureof Terragreen (oil dry US special type II R, granular clay mineralAttapulgite;>0.125 mm) and Loess (3:1; pH-KCl 6.2; organic carbon0.3%; easily and slowly available P (‘P-Morgan’ (=P-Na-acetate) andP-citrate, respectively) 2.6 mg kg�1 and 533 mg kg�1; available K(Na-acetate) 350 mg kg�1) was used as substrate for all microcosms;after application of the inocula, three 2-weeks old AM fungi-freeseedlings of each of the three plant species were planted permicrocosm in close vicinity of the inocula. An automated wateringsystem (Tropf-Blumat, Weninger GmbH, A-6410 Telfs, Austria)maintaining the water holding capacity at about 80% throughout theexperiment, prevented water stress and reduced the risk of crosscontamination between the AMF communities by water splashes.The microcosm experiment was conducted under near natural lightand temperature conditions in a greenhouse except that thetemperature was not allowed to fall below the freezing point:nevertheless, the climatic conditions of the Upper Rhine Valley (i.e.the region of origin of the AMF inocula) were approximately met alsoduring the winter periods in the greenhouse as the coldest month inthe Upper Rhine valley is January with about +1 8C averagetemperature and between 3 and 8 ice-days only. The experimentended in December 2003 after 32 months, covering three completevegetative periods. The plant consortium was cut 3 cm above theground four times per season, usually 3–5 days before substratesampling. Plant growth was season-dependent ceasing in eachwinter between December and February, restarted in the middle ofMarch in the second and third vegetation periods without visibledifferences between the 3 years. Fertilization was not necessary,profiting from the nitrogen fixing T. pratense that had been includedin the host plant consortium. Since the substrate mixture was wellsupplied with P, K (see above), and with Mg, Ca and micronutrients(data not shown), the plants never showed any nutrient deficiencysymptoms during 32 months.

2.4. Sampling of newly formed spores in the microcosms

During the first winter we checked for spore formation andfound no spore reproduction between the beginning of Decemberand the beginning of April. Therefore, throughout the experiment,we sampled at four dates per year in bimonthly intervals beginningin June and ending in December (i.e. in spring (June), summer(August), early autumn (October, and early winter (December)) forextracting AMF roots and spores from each microcosm. Each time,two soil cores (each 15 cm3, sampling depth 10 cm) were takenvertically at the same location in the microcosms. Immediatelyafter sampling, the holes were refilled with the same sterilizedsubstrate used to set up the microcosms but only up to 5 mmbelow substrate surface to allow re-sampling at exactly the sameposition each time. Thus, the samples contained only newlyformed spores produced during the 2 months passed after previoussampling or during the period between the last sampling at the endof the season and the first sampling at the beginning of June. At theend of each vegetative period (i.e. after 8, 20 and 32 months of trapculturing) two additional soil cores per microcosm were taken at aseparate location. The substrate was replaced in the manner asdescribed above. Thus, these samples contained the sporepopulation produced within one season. Finally, at the end ofthe microcosm experiment after 32 months, a third pair of soilcores per microcosm was taken at locations that had not beensampled previously.

2.5. AMF spore isolation and identification

AMF spores were extracted from the soil samples by wet sievingand sucrose density gradient centrifugation as described pre-

viously (Oehl et al., 2003a). Spores, spore clusters and sporocarpswere picked without pre-selection and mounted together onmicroscope slides using polyvinyl–lactic acid–glycerol (PLVG) orPLVG mixed 1:1 (v/v) with Melzer’s reagent (Brundrett et al., 1994)as semi-permanent mounts. The slides were examined system-atically under a stereo compound microscope (Zeiss, Axioplan) atup to 630-fold magnification to identify all morphologicallydistinct AMF spore types present. Identifications were based oncurrent species descriptions and identification manuals (Schenckand Perez, 1990; International Culture Collection of Vesicular andVesicular-Arbuscular Endomycorrhizal Fungi [http://invam.caf.w-vu.edu/Myc_Info/Taxonomy/species.htm]). If not marked other-wise (see legend in Table 1), a species was recorded as beingpresent at a given field site when it was represented by severalspores in at least three of the four field plot replicates.

Relative abundance of spores of the AMF species identified wasestimated on the microscope slides for at least 100 randomlyselected spores per replicate plot. A species was judged as‘dominant’ when it comprised at least 15–20% of all sporesidentified per sampling period.

2.6. Data analyses

The significance of differences between the cumulative AMFspecies numbers after 32 months in the microcosms represent-ing the different field sites was tested using Fisher’s leastsignificant difference (LSD) at P < 0.05 after a one-way analysisof variance (ANOVA). Linear regression was applied to check forsignificant increase of cumulative species numbers between thelast sampling date of the second season (20 months) and the endof the experiment after the third season (32 months). At eachsampling date species numbers, and relative spore abundance ofthe 20 most dominantly sporulating species in the microcosmsare given as mean of 4 replicates per sites, with standarddeviation and standard error, respectively. As multivariateanalysis a canonical cluster analysis (CCA) was performed aftera preliminary Detrended Correspondence Analysis (DCA)(Canoco 4.5) on the mean relative spore abundances per species,site and date. These calculations focused (i) solely on time (12sampling dates in 3 vegetation periods) and season (spring,summer, early autumn and early winter), (ii) including also landuse intensity, and (iii) several other environmental parameters(soil type, soil pH, organic carbon, available P) and sitespecificity.

3. Results

3.1. AMF species found in the microcosms

Overall, we could distinguish thirty-nine AMF species withinseven genera on the basis of spore morphology (Table 1). Thirty-three of these spore morphotypes could be assigned unequi-vocally to known species, namely 25 species of Glomus (G), fourspecies of Acaulospora (A), and one species each of Entrophospora

(E; Sieverding and Oehl, 2006), Paraglomus (P), Archaeospora (Ar),Ambispora (Am) (Spain et al., 2006; Walker, 2008) and Scutellos-

pora (S), and one species of the new genus Cetraspora (Oehl et al.,2008). Six distinct, regularly occurring spore morphotypes couldnot be clearly identified at the species level and might representhitherto undescribed species (Table 1). Meanwhile two formerlyundescribed species (Glomus sp. strains BR2 and BR4; Oehl et al.,2003a) have been described as G. aureum and G. badium,respectively (Oehl et al., 2003b, 2005a), and are listed underthese names. Another morphotype formerly supposed to beundescribed (Glomus sp. strain BR3; Oehl et al., 2003a) was foundto be G. constrictum (see Table 1). No representatives from the

Table 1Time of first appearance (month) of spores of AMF species in the course of three vegetative periods (32 months) in microcosms set up with AMF inocula from the various field

sites, ordered according to ‘‘eco-groups’’ (see Fig. 3A–C).

AMF species or strain Field site used as source of AMF inoculuma

Grassland Arable land

W V G Crop rotation Monocropping (maize)

O L F S R

Glomus eco-group 1 A (early successional stage, early in first season, occurring in all habitats)

Glomus mosseae 4 4 4 2 2 2 2 2

Glomus etunicatum 8 8 6 2 2 4 4 4

Glomus geosporum 26 8 8 2 2 20b 16 4

Glomus lamellosum 6 4 6 8 4 4

Glomus diaphanum 18 8 18 6 18 2 4 8

Glomus albidum 16 6 8 4 6 8 8 6

Glomus aggregatum 28 14 6 4 26 6 4 6

Glomus eco-group 1 B (early successional stage, not in grassland)

Glomus caledonium 4 2 2 2

Glomus eco-group 2 (intermediate successional stage, late in first season)

Glomus constrictumc 6d 6d 8d 6d 16 32b 14 14

Glomus fasciculatum 6 6d 6d 8d 8d 8d 16 18

Glomus sp. strain BR11e 8d 8d 6d 8d 14 20 18 18

Glomus versiforme 14 14 6d 6d 6d 26 14 14

Glomus eco-group 3A (late successional stage, occurring in most habitats)

Glomus sp. strain BR9 16 20 8d 6d 18

Glomus invermaium 26 28 18 16 16 18 30

Glomus aureumf 8d 30 18 16 8d 8d 16 28

Glomus sp. strain BR12g 18 16 16 16 16 14

Glomus intraradices 26b 32b 8b 14b 20b 14b 16

Glomus eco-group 3B (late successional stage, not in monocropping systems)

Glomus macrocarpum 28 28 26

Glomus microcarpum 26 20 20 8d 26

Glomus badiumh 32b 8b 20

Glomus sp. strain BR13 32 28 30

Glomus eco-group 3C (late successional stage, only in grassland)

Glomus sinuosum 26

Glomus mortonii 32b

Glomus heterosporum 32b 20 30

Acaulospora eco-group (main sporulation early in the season)

Acaulospora paulinae 16 6d 8d 14b

Acaulospora longula 14 16 8d 14

Acaulospora thomii 16 8d 14

Acaulospora laevis 16 14

Archaeospora trappei 14 16 8 6 4 8 14

Ambispora fennicai 16

Scutellospora eco-group (sporulation late in the season, not in monocropping)

Scutellospora calospora 8 8 20

Scutellospora pellucida 20b 32

Not attributed to any ecological group as only found in few specimens

Entrophospora infrequens 8d 26

Glomus sp. strain BR5 8 6

Glomus sp. strain BR15j 14b

Paraglomus occultum 20b

Glomus clarum 20b

Glomus pustulatum 30b

Glomus coronatum 32b

a The AMF inocula in the microcosms were derived from eight field sites (W–R) differing in agricultural land use (Oehl et al., 2003a). Input and management intensity

increasing from left to right.b Found only in <10 specimens during the whole experiment (0–32 months).c Glomus sp. strain BR3 (see Oehl et al., 2003a) is now considered as synonymous with G. constrictum.d During the first vegetative period (0–8 months) only found in <5 specimens.e Resembles G. arborense.f G. aureum presented as G. sp. strain BR2 in Oehl et al. (2003a) was recently described from the study area (Oehl et al., 2003b).g Resembles G. aurantium.h G. badium presented as G. sp. strain BR4 in Oehl et al. (2003a) was recently described from the study area (Oehl et al., 2005a).i Ambispora fennica (Walker et al., 2007) was called Acaulospora sp. BR13 in Oehl et al. (2004).j Resembles G. fuegianum.

F. Oehl et al. / Agriculture, Ecosystems and Environment 134 (2009) 257–268260

genus Gigaspora nor from Pacispora were observed in the presentstudy, although known from the study area. Typically, about 75%of the spores in a given sample could be assigned to one of the 39species, except at a few sampling dates when, due to an

exceptionally high proportion of immature spores, the percentagewas slightly lower.

Whilst five of the non-inoculated control microcosms did notcontain any AMF spores until the end of the experiment three of

Fig. 2. Cumulative total number of AMF species detected by spore morphotyping in

samples taken at regular intervals over three growth seasons (32 months) from

microcosms that had been inoculated with soil from field sites (W, V, G, O, L, S, F, R)

differing in agricultural land-use. Thin lines show linear regressions over the last

F. Oehl et al. / Agriculture, Ecosystems and Environment 134 (2009) 257–268 261

them contained spores of a single AMF species, namely of G.

intraradices, indicating that this species was the only potentialextraneous contaminant in the study. G. intraradices spores werefound in all microcosms that had been inoculated with soil samplesfrom the field sites, and the morphotype variability of the G.

intraradices spores in these microcosms was higher than in thethree infested control microcosms. We therefore concluded thatthe G. intraradices spores in the inoculated microcosms most likelyrepresent genuine species from the field sites and were not casualgreenhouse contaminants.

3.2. Number of sporulating AMF species over the three seasons of

sampling

In the microcosms set up with the inocula from the maizemonocropping systems, a total of four to five AMF species werealready present in the 4-months-samples, and six to eight speciesin the 8-month samples. Beyond this point the number of speciesreached an asymptote for the remainder of the experiment (Fig. 1).In the microcosms set up with the inocula from arable fields undercrop rotation, the numbers of species initially were similar to thatin monoculture (five to six in 4-month samples) and also reachedan apparent asymptote in 8-month samples, but the numbers of

Fig. 1. Number of AMF species detected by spore morphotyping in samples taken at

regular intervals over three growth seasons (32 months) from microcosms that had

been inoculated with soil from field sites (W, V, G, O, L, S, F, R) differing in

agricultural land-use. (A) Average species numbers (four replicates per site) with

standard deviation and (B) total species numbers detected at each sampling date.

Soil cores were taken each time at the same location in the microcosms to ascertain

that exclusively spores newly formed during the sampling intervals were seized.

five sampling dates; slopes are steeper for AMF communities from the grasslands

(W = 0.64*; V = 0.40*; G = 0.48*) than for AMF communities from the low- to

medium-input arable lands (O = 0.10, L = 0.40*) and from the high-input, mono-

cropped arable lands (F = 0.14*, S = 0.17*, R = 0.10); * denotes a significantly positive

slope (linear regression at the 5% significance level).

species were higher (12–18). In contrast, in the microcosms set upwith grassland inocula, one to two species were found in the 4-month samples, but the numbers of species per sampling timecontinued to increase and reached similar values as in the samplesfrom the crop-rotation microcosms in the third season (11–14 AMFspecies).

The same data set is presented in a cumulative manner(Fig. 2), i.e., at each time point, the total numbers of AMF speciesfound in the various sets of microcosms at this time point plusall previous ones were calculated. During the first season, for allthe sets, the cumulative numbers of species were essentially thesame as the numbers of species represented by newly formedspores at the corresponding sampling time (see Fig. 1B).However, for all the sets of microcosms, the cumulativenumbers of species continue to increase during the secondseason (Fig. 2) despite the apparent asymptote reached in thenumber of species at a given sampling time (Fig. 1). Thisrepresents successional sporulation: some AMF species nolonger sporulated while others began to sporulate only at alater stage of microcosm development. In four out of five groupsof microcosms set up with inocula from the five arable lands(both monocropping and crop rotation), the cumulative num-bers of species did not increase further, as indicated by veryshallow, non-significant slopes (0.10–0.17) of the linear regres-sion over the last five sampling times; in contrast, thecumulative numbers of species still increased significantly inall sets of microcosms set up with inocula from grasslands(slopes 0.40–0.64), indicating that in this case the successioncontinued up to three seasons (Fig. 2).

3.3. Temporal sporulation patterns: ecological groups of AM fungi

To further explore the concept of temporal sporulationdynamics, an attempt was made to group the different AMFspecies according to their earliest appearance in the microcosms(Table 1) and their later disappearance among newly formedspores (Fig. 3; summarily shown in Fig. 4). Species of the first

Fig. 3. Relative AMF spore abundance at different sampling dates (average of four replicates per site and sampling date, with standard error) for the 20 most abundantly

sporulating species found in the microcosms set up with AMF inocula from eight field sites (W–R) differing in agricultural land use. Symbols for species attributed to the

Glomus eco-groups 1, 2 and 3 and Acaulospora eco-group are presented in white, gray, black and dark gray fill colors, respectively.

F. Oehl et al. / Agriculture, Ecosystems and Environment 134 (2009) 257–268262

Glomus group (in this study called Glomus eco-group 1) sporulatedvery early, particularly in the microcosms set up with inocula fromarable lands, and were usually recorded for the first time withinthe first 2–4 months (Table 1; Fig. 3). Most of these AM fungi

(designated eco-group 1A) also sporulated in microcosms set upwith inocula from grasslands; however, they were generallyrecorded for the first time somewhat later, after 4–8 months(Table 1). Remarkably, the spore density of species of this group

Fig. 4. Semi-quantitative analysis of the spore population of AMF species in the

different sets of microcosms, in the course of three growth seasons (32 months).

The species were attributed to ‘‘ecological groups’’ (see Table 1). Data for

microcosms set up with inocula from: (A) intensively managed arable lands

under maize monocropping (mean of sites F, S and R), (B) low- to medium input

arable lands under rotation (mean of O and L) or (C) extensively managed

grassland soils (mean of W, V and G). Thickness of stripes at a given time

period show the estimated relative abundance of spores belonging to

different eco-groups of AMF species, with four steps of relative abundance:

.

F. Oehl et al. / Agriculture, Ecosystems and Environment 134 (2009) 257–268 263

decreased dramatically in all the various sets of microcosms duringthe third season of the experiment (Figs. 3 and 4A–C). One memberof this group, Glomus caledonium, which sporulated particularlyquickly in the microcosms set up with inocula from arable lands,was never recovered from microcosms derived from the grasslands

(Table 1), and it was therefore put as a single member in Glomus

eco-group 1B.Glomus species of the second group (called eco-group 2) usually

began sporulating in the microcosms at the second half of the firstvegetative period, i.e. after 6–8 months (Table 1). They werepresent in almost all microcosms, and they continued to sporulateabundantly during the second and third season (Figs. 3 and 4A–C).Interestingly, species in this group appeared first after 14 monthsin microcosms set up with inocula from monocropping systems.

Glomus species of the third group (eco-group 3) usuallyrequired more than one vegetative period to start with sporulation,and they may be considered to represent a late stage of AMFsuccession (Table 1). Species of this group sporulated mainlyduring the second half of the second vegetative period (Table 1;Figs. 3 and 4). Some species of Glomus eco-group 3 occurred inmicrocosms derived from all the various field sites (designatedeco-group 3A), but they often appeared relatively rarely in themicrocosms derived from the most intensively managed sites F, Sand R (Table 1). Other species of Glomus eco-group 3 (designatedeco-group 3B) were found in all sets of microcosms except thosederived from the most intensively managed sites (Table 1; Figs. 3and 4). The occurrence of a further set of species in group 3(designated eco-group 3C) was even more restricted; theyoccurred only in microcosms set up with inocula from thegrasslands; they formed spores only during the third vegetativeperiod (Table 1; Fig. 4C).

All species of Acaulospora, Archaeospora and Ambispora (desig-nated Acaulospora eco-group) recorded in this study sporulatedearly in the growing seasons in all the microcosms in which theywere found, and sporulation decreased drastically towards the endof the growing seasons (Figs. 3 and 4A–C). Spores of three otherspecies recorded, S. calospora and C. pellucida (designatedScutellospora eco-group) and G. badium (with low sporulationfrequency), were only found in the samples taken at the lastsampling date in each of the 3 years (Table 1). Like the Acaulospora

spp., these species were mainly found in the microcosmsrepresenting the less intensively managed arable lands or thegrasslands (Table 1). The remaining species listed in Table 1 weredetected only infrequently and, thus could not readily be assignedto any of the ecological groups. The majority of these speciesbelonged to the genus Glomus (Table 1).

In the ordination diagram, focusing only on succession (time)and seasonality of sporulation as environmental variables, allspecies of the Glomus eco-groups 1, 2 and 3 (with exception of G.

badium and G. macrocarpum) grouped along the time vector andformed three clearly separated clusters (Fig. 5A). S. calospora and C.

pellucida clustered together and were, like G. badium, closer relatedwith the season than the time vector, while the Acaulospora speciesand Ar. trappei clustered in inverse position to both, the seasonvector and those three species (Fig. 5A). When the factor land useintensity was included in the analysis (Fig. 5B), the three Glomus

eco-groups sub-divided along the land use intensity vector. Thesub-division of Glomus group 3 was most obvious, with sub-clusters of G. intraradices and Glomus sp. BR12 of eco-group 3A, G.

macrocarpum and G. microcarpum of eco-group 3B and G. sinuosum

and G. heterosporum of eco-group 3C. The Acaulospora speciesclustered in the inverse direction of the land use intensity vector.When additionally other environmental variables such as soil type,pH, organic C, available P and the site effect were considered, thespecies groups separated less, and several vectors seemed to besimilar or inversely directed (data not shown).

3.4. Dominance of AMF species at different sampling times

In the first 4 months of the experiment, as expected, spores ofGlomus eco-group 1 (particularly G. mosseae, G. etunicatum and G.

Fig. 5. Ordination diagrams showing the effect of time and season on the relative spore

abundance of different AMF species, A. without, and B. with consideration of the land

use intensity as additional environmental variable. Canonical Corresponding Analyses

(CCA) were performed on spore data of twelve sampling dates (2, 4, 6, 8, 14, 16, 18, 20,

26, 28, 30, 32), four seasons (0 = spring, 1 = summer, 2 = early autumn, 3 = early

winter sampling) and four levels of land use intensity (0 = semi-natural grasslands,

1 = bio-organic, crop rotation, 2 = conventional Integrated Production, crop rotation,

3 = conventional, maize monocropping). The eigenvalues are 0.473, 0.074, 0.514 and

0.368, and 0.515, 0.166, 0.073 and 0.504 in A and B, respectively.The diagrams account

for 100% and 90.3% of the variance of the sporulation data, respectively. Symbols for

species attributed to the Glomus eco-groups 1, 2 and 3 and the Acaulospora and

Scutellospora eco-groups are presented in white, gray, black and dark gray fill colors,

respectively. Full species names see Table 1.

Fig. 6. Cumulative number of AMF species (average of four replicates per site, with

standard deviation) found in the microcosms representing the field sites (W, V, G, O,

L, F, S, R) differing in agricultural land-use. The microcosms were sampled regularly

during three seasons (32 months). Soil cores were taken each time at the same

location in the microcosms to ascertain that exclusively spores newly formed

during the sampling intervals were seized. Non-significant differences between

field sites are indicated by identical letters above the bars and were determined by

using Fisher’s least significant difference at the 5% level after one-way ANOVA.

F. Oehl et al. / Agriculture, Ecosystems and Environment 134 (2009) 257–268264

caledonium) were dominant in all microcosms (Figs. 3 and 4). In thesecond half of the first growing season these fungi remaineddominant, but some species of eco-group 2 became abundant aswell, particularly in the microcosms set up with inocula from thegrasslands (G. constrictum and G. fasciculatum). In the first half ofthe second season, G. mosseae and G. etunicatum disappeared asdominant species in most types of microcosms, while Archaeospora

trappei became a dominant species in three types of microcosms.Interestingly, in microcosms set up with inoculum from the croprotation system with organic farming, Acaulospora species weredominant in this period (Fig. 3). In the second half of the secondseason, G. constrictum (eco-group 2) became the most dominantspecies in the majority of the microcosms, but some species of eco-

group 3 (e.g. G. invermaium and G. aureum) became dominant aswell. In the first and second half of the third season, patterns ofdominant species in a given type of microcosm were generallysimilar to the patterns of the same type of microcosm in the firstand second half of the second season, indicating that a kind ofequilibrium was reached. However, in microcosms derived fromgrassland ecosystems, members of the eco-groups 2 and 3 (e.g. G.

aureum) became more prominent.

3.5. Number of sporulating AMF species detected when samples were

taken only at the end of each season or only once at the end of the

experiment

When samples were taken in the microcosms at yearlyintervals, the cumulative numbers of species were somewhatreduced compared to the bimonthly analysis (Table 2). Never-theless, the cumulative numbers increased from year to year,indicating that successional sporulation occurred under theseconditions as well. Samples taken in the microcosms at previouslyunsampled locations only at the end of the experiment (Table 2)contained similar numbers of AMF species as the samplesrepresenting the newly formed spores within the two last monthsof the experiment (Fig. 1, data points at 32 months). Interestingly,species of Glomus eco-group 3A (e.g. G. invermaium and G. aureum)were regularly sporulating in the second and third season in themicrocosms set up with inocula from the monocropping agro-ecosystems when sampling was performed in bimonthly intervalsor at the end of each vegetative period, but they rarely occurred inthe same microcosms in the samples taken after 32 months fromuntouched locations (data not shown).

3.6. Number of AMF species found in different agroecosystems

The AMF species numbers were significantly higher in themicrocosms representing the grasslands and the arable landssubjected to crop rotation than in the microcosms representing themaize monocropping systems (Table 3; Fig. 6). Remarkably, thearable land under organic farming (site O) produced the highestAMF species richness, even higher than the grasslands. Also theconventionally arable land L had higher species numbers than twoof the three grasslands (Table 3; Fig. 6).

Table 3Synopsis of the number of AMF spp. found directly in field samples (Oehl et al., 2003a) and subsequently in the microcosms set up with inocula from these field sites (W–R).

Field site

Grassland Arable land

W V G Crop rotation Monocropping (maize)

O L F S R

Number of AMF species found in the field samplesa 23 24 19 18 13 10 9 8

Number of AMF species (cumulative, total) found in the

microcosms during 32 months (Table 1 and Fig. 2)

21 24 26 31 25 19 18 13

Number of AMF species found in the field and also in the

microcosms during 32 months

15 17 16 18 11 10 9 7

Number of AMF species not found in the field but in the

microcosms during 32 months

6 7 10 13 14 9 9 6

a Species numbers presented previously (Oehl et al., 2003a) are here corrected as Glomus sp. strain BR3 and G. constrictum (Oehl et al., 2003a) are now considered as a single

AMF species.

Table 2Cumulative total numbers of AMF species found after one to three seasons (upon sampling once per season) in the microcosms set up with AMF inocula from the various field

sites, compared to the numbers found after three seasons (upon sampling only once at the end).

Sampling time Field site used as source of AMF inoculuma

Grassland Arable land

W V G Crop rotation Monocropping (maize)

O L F S R

Cumulative total numbers of AMF over three seasons (sampling once per season, at the same locations in the microcosms each time)

After one season (8 months) 6 9 8 13 11 6 6 5

After two seasons (20 months) 9 12 11 18 19 9 9 7

After three seasons (32 months) 16 15 18 25 24 12 11 10

Number of AMF species after three seasons (sampling once at the end of the experiment at a location that had not been touched before)

After three seasons (32 months) 10 10 12 19 14 9 10 7

a The AMF inocula in the microcosms were derived from the eight field sites (W–R) differing in agricultural land use (Oehl et al., 2003a). Input and management intensity

increasing from left to right.

F. Oehl et al. / Agriculture, Ecosystems and Environment 134 (2009) 257–268 265

4. Discussion

4.1. Number of AMF species in the microcosms as compared to those

found at the field sites

In our previous work, we found that after short-term (8months) the number of AMF species in the microcosms set up withinocula from the arable lands (monocropping or crop rotation)reflected the species diversity of the field samples very well;however, to our surprise, the microcosms set up with AMF inoculafrom the low-input grassland sites contained much lower numbersof AMF species than the respective field samples (Oehl et al.,2003a). The overall result of the present study (Table 3), now leadsto a quite different conclusion: In the microcosms set up withinocula from the low-input grasslands, the cumulative numbers ofspecies in the microcosms matches the numbers of AMF species inthe field samples, although some of the typical ‘AM fungalspecialists’ (Oehl et al., 2003a,b) were not recovered even in theselong-term microcosms. However, in the microcosms set up withinocula from the arable lands, the cumulative numbers of speciessporulating at some time during the three seasons are up to twotimes higher than that found directly in the field samples (Table 3;Fig. 5). Hence, the previous field analysis (Oehl et al., 2003a)considerably underestimated the AMF species richness in arablelands.

Comparing the total and average cumulative numbers of AMFspecies found in the various sets of microcosms (Figs. 2 and 5,respectively), the microcosms representing high-input arable land(i.e. maize monocropping) contained only about half as manyspecies as the ones representing arable land under crop rotation orgrassland. Since maize monocropping represents a strong dis-turbance (bare soil without any vegetation from about November

till April in the region investigated), AMF species may survive onlyif they can complete their life cycle and form spores within thevegetative period of maize—a condition that selects strongly forAMF species of the Glomus groups 1 and 2. Indeed, a recent study ofsite ‘‘R’’ using molecular tools clearly indicated a complete absenceof Acaulospora and Scutellospora species in the roots under theseconditions (Hijri et al., 2006), which, however, might have alsobeen dependent on the soil pH (Oehl, unpublished). The absence ofone or several specific AMF eco-groups in the intensively managedarable lands suggests loss of biodiversity and might represent alsoa reduced ecosystem function if the sites are converted to moresustainable, low-input arable lands practicing crop rotation or tospecies rich grasslands. However, the functional role of differentecological AMF groups in different ecosystems is poorly under-stood so far. Interestingly, cumulative AMF species richness wasvery high in the microcosms representing the arable land subjectedto crop rotation, particularly in site O that was fertilized accordingto a bio-organic scheme (Mader et al., 2002), indicating that theregular change of crop plants, use of cover crops and organicfertilizers may help to maintain a rich community of AMF species.In the case of microcosms representing grassland, the cumulativenumbers of AMF species after 3 years may not represent the fullextent of diversity, as indicated by the fact that the regression lineover the last five sampling points was clearly positive (Fig. 2).

4.2. Successional sporulation dynamics

One intriguing aspect of the present work is the finding thatAMF consortia derived from all field sites exhibited strongtemporal sporulation dynamics. This is in accordance with veryrecently published results of Liu et al. (2009) that found temporalsporulation dynamics in a chronosequence of Caragana korshinskii

F. Oehl et al. / Agriculture, Ecosystems and Environment 134 (2009) 257–268266

plantations, however, with a rather low overall AMF speciesrichness and very low AMF genus richness at those sites. With ourlong-term microcosm systems that included AMF communities ofboth high species and genus richness derived from three differentagroecosystems, we could go beyond such observations andidentify several AMF eco-groups with distinct sporulation pat-terns. Independently of the origin of the source of the AMFinoculum, all microcosms initially showed a prevalence of Glomus

spores of eco-group 1. The species represented by this group wereclearly dominant among the AMF spores in field samples from themono-cropped sites and were much less prevalent or even absentin the field samples from the less intensively managed sites (Oehlet al., 2003a). Thus, high-intensity agriculture such as maizemonocropping, with its strong disturbance upon harvest, appearsto strongly select for AM fungi that form spores quickly. It is worthnoting that all of these species, with the exception of G. caledonium

(Glomus eco-group 1B), were also found in microcosms set up withinocula from the grasslands; however, they became apparent onlylate in the first season or even only in the second or third season.Either, the ecotypes of these fungi occurring in grassland formspores only late and rarely, or, according to Gazey et al. (1992),their inoculum potential had first to increase to a level that enabledspore formation.

Interestingly, spores of Glomus eco-group 2 tended to appearearlier in microcosms set up with inocula from low-input sites thanhigh-input sites, reflecting either ecotype differentiation as well orsignificantly decreased inoculum potential of these species in thehigh-input soils. The latest stage of sporulation is represented byGlomus species of eco-group 3. None of these AM fungi hadpreviously been recovered from field samples from the high-inputmonocropping systems (Oehl et al., 2003a). Nevertheless, some ofthem, such as G. invermaium and G. aureum, established themselveswell in the microcosms set up with inocula from these high-inputmonocropping systems and, remarkably, had been detected in fieldsamples of undisturbed sub-soils of these systems (Oehl et al.,2005b). Obviously, there is a relationship between some aspects oflife history (here how long it takes to sporulate) and the occurrenceof Glomus species in low-input and high-input arable lands.However, our observations show that the loss of AMF diversity byintensified agricultural practices is less strong than assumed fromfield sample analyses solely (Oehl et al., 2003a) suggesting that thereinoculum potential of specific AMF species has remained althoughwe had not detected their spores in the field samples.

4.3. Seasonal sporulation dynamics

With exception of scarcely sporulating G. badium (Table 2 andFig. 5A), there was no indication for seasonal sporulation amongmembers of the genus Glomus in the microcosms, whilst membersof some other genera displayed a strongly seasonal sporulationpattern. All the Acaulospora species recorded as well as Archae-

ospora trappei and Ambispora fennica had a peak of sporulationearly in the season (April–July), and Scutellospora calospora andCetraspora pellucida sporulated only in late fall (between Octoberand December of all 3 years). Spores from these fungi were rarelyobserved in samples from the field site, except for field site O (Oehlet al., 2003a). A similar type of seasonality was found in a fieldstudy of blue bells (Hyacinthoides non-scripta) in England (Merry-wheather and Fitter, 1998b): During spring time, arbuscules ofAcaulospora dominated in the roots of blue bell, and spores of A.

koskei in the surrounding soils, while in the same plant duringautumn, arbuscules that were attributed to Scutellospora domi-nated the roots and S. dipurpurescens spores the surrounding soils.There have been a few studies describing seasonal sporulationpatterns in these groups of fungi: Pringle and Bever (2002)reported for a grassland in North Carolina that especially A.

colossica and Gigaspora gigantea had different and contrastingsporulation seasonalities. Gemma and Koske (1988) and Lee andKoske (1994) found peak sporulation for Gigaspora gigantea duringfall in a coastal sand dune ecosystem in Rhode Island. Usingmolecular tools, Daniell et al. (2001) and especially Hijri et al.(2006) more frequently found Acaulospora sequences in field rootssampled in spring to early summer, while sequences belonging tothis genus were not or in low frequency obtained from rootssampled later in the season. The later authors referred to theunpublished results of the present study.

The fact that Glomus spp. generally had rather a successional,while many species of several non-Glomus genera had a stronglyseasonal sporulation pattern might be related to their different lifehistory traits: the majority of Glomus spp. may germinate fromspores as well as from mycorrhizal roots (especially vesicles),while Acaulospora and especially Scutellospora and related genera(e.g. Cetraspora and Gigaspora) are believed to germinate pre-ferably or exclusively from spores after a dormancy period (e.g.Gazey et al., 1993; Brundrett et al., 1999). The divergingsporulation patterns of Acaulospora and Scutellospora eco-groupsmight be due to the ability of Acaulospora spp. to be physiologicallyactive already in the cool season and to last through the summer asa dormant spore (Pringle and Bever, 2002). These observations arein accordance with our recent findings (Oehl et al., 2006; Spainet al., 2006) that spores of Acaulospora, and also Ambispora spp. aremore abundant in short-seasoned high mountainous and alpineelevations of the Swiss Alps than in the warmer Central Europeanlowlands, especially when compared to Scutellospora, Cetraspora

and Gigaspora spp. (Oehl et al., unpublished).

4.4. Small scale disturbances affecting sporulation dynamics

The sampling scheme developed in this work, in which the samelocation in the microcosm was sampled at bimonthly intervals, byreplacing the sampled spore-containing soil with fresh sterile soil,has the advantage to monitor only the spores newly formed at given2-month periods and allows a detailed analysis of the dynamics ofspore formation. However, it also means a small scale disturbanceand offers a new ‘‘empty space’’ for development of roots,extraradical mycelium and spore initials. We do not know to whatextent this interference influences the sporulation pattern of theAMF community. A comparison with samples taken only once a yearclearly indicates that the bimonthly sampling yielded considerablylarger cumulative numbers of AMF species (compare Tables 2 and 3).In part, this might have to do simply with the fact that four timesmore soil volume was inspected. However, it may also be that thefrequent introduction of ‘‘empty space’’ in an established microcosmmay be conducive to sporulation of some AMF species, particularlyto those of Glomus eco-group 3. It is worth noting that at least for themicrocosms set up with inocula from grasslands, samples taken atthe end of the experiment from previously untouched locationscontained spores of only ca. 10 AMF species, less than half of thecumulative number of AMF species found in the bimonthly analysis.Small-scale disturbances such as the one introduced by oursampling scheme may also occur in field situations (e.g. causedby the soil macro-fauna such as earthworms) and may be importantfor sporulation induction of AM fungi. Earthworms are believed to beimportant vectors of AM fungi (Rabadin and Stinner, 1989; Redelland Spain, 1991), and Gange (1993) showed that earthwormmaterial contained higher numbers of spores and infectivepropagules than nearby field soil.

4.5. ‘Long-term memory’ of AMF populations

Our microcosms representing model permanent grasslandswere kept on fully identical conditions for 3 years, except for the

F. Oehl et al. / Agriculture, Ecosystems and Environment 134 (2009) 257–268 267

AMF inoculum that came from different field sites. It isintriguing to ask whether the AMF populations in all ofthese microcosms converge towards a sort of climax commu-nity, or whether they maintain some individual identitiesreflecting the source of the inocula. Considering the dominantAMF species in the second half of the third season (Fig. 3,months 28–32), there are some indications for convergence: Inseven out of the eight microcosms, G. constrictum (the mostabundant AMF species in Polish grasslands according toBłaszkowski (1993) and found in all the field samples, althoughat a relatively low spore density of 1–10%; Oehl et al., 2003a)was the most dominant or second-most dominant speciesamong the population of new spores, indicating that itflourished best under the conditions of the microcosms.However, even more evidently, the microcosms retained someidentity reflecting the origin of the inoculum. For example, G.

mosseae, a dominant species in field samples from the high-input monocropping system, initially was dominant in all themicrocosms, but almost completely disappeared again in all themicrocosms except for the one representing the monocroppingsystem with the highest input, site R. On the other hand, G.

aureum, which in the field study was found almost exclusively insamples from the grasslands, appeared after the first season inmicrocosms representing all eight field sites but became adominant species at the end of the experiment only inmicrocosms representing the grassland AMF community. Thus,surprisingly, the AMF populations in the microcosms set up withinocula from the different field sites initially ‘‘lose identity’’ butlater re-gain it, indicating a sort of long-term memory of theAMF population even under fully equal conditions.

5. Conclusions

The objectives of this study were to assess temporalsporulation dynamics of AMF communities in microcosmsderived from different agroecosystems, and to possibly recog-nize different ecological AMF groups. Analyses of newly formedspores in the microcosms according to our novel bimonthlylocalized sampling scheme revealed clear successional andseasonal dynamics of AMF spore formation, implying differentlife strategies of different AMF species. Thus, different ecologicalAMF groups could be defined on the bases of diverging temporalsporulation dynamics. Our data confirm that under low-inputfarming, especially under organic farming, a high AMF biodi-versity can be established. Remarkably, the data further suggestthat in such farming systems the diversity might be evenhigher than under natural grasslands. On the other hand, AMfungi, belonging to specific ecological AMF groups, disappearedunder high-input maize monocropping. The loss of ecologicalAMF groups, however, may indicate a loss of soil function andsoil quality, which may become important when land use ischanged to more sustainable (and diversified) agriculturalsystems.

Acknowledgments

We thank Giacomo Busco for technical assistance and thestudents Robert Bosch, Rolf Spirig, Andreas Keller and SueFurler for spore isolation. We are grateful for Dr. Jurg Hiltbrunner(Agroscope Reckenholz-Tanikon Research Station ART, Zurich,Switzerland) for statistical advice. Dr. Dirk Redecker is acknowl-edged for critically reading the manuscript (University of Basel,Switzerland). This study was supported by the Swiss Agency forDevelopment and Cooperation (SDC) in the frame of the Indo-SwissCollaboration in Biotechnology (ISCB) program and by the SwissNational Science Foundation.

References

An, Z.Q., Guo, B.Z., Hendrix, W., 1993. Populations of spores and propagules ofmycorrhizal fungi in relation to the life cycles of tall fescue and tobacco. SoilBiol. Biochem. 25, 813–817.

Beauchamp, V.S., Stromberg, J.C., Stutz, J.C., 2006. Arbuscular mycorrhizal fungiassociated with Populus-Salix stands in a semiarid riparian ecosystem. NewPhytol. 170, 369–380.

Bentivenga, S.P., Hetrick, B.A., 1992. Seasonal and temperature effects on mycor-rhizal activity and dependence of cool and warm-season tallgrass prairiegrasses. Can. J. Bot. 70, 1596–1602.

Bever, J.D., Schultz, P.A., Pringle, A., Morton, J.B., 2001. Arbuscular mycorrhizal fungi:more diverse than meets the eye, and the ecological tale of why. Bioscience 51,923–931.

Błaszkowski, J., 1993. Comparative studies on the occurrence of arbuscular fungiand mycorrhizae (Glomales) in cultivated and uncultivated soils of Poland. ActaMycologica 28, 93–140.

Błaszkowski, J., 1994. Arbuscular fungi and mycorrhizae (Glomales) of the HelPeninsula, Poland. Mycorrhiza 5, 71–88.

Błaszkowski, J., Tadych, M., 2000. Succession of arbuscular mycorrhizal fungi in adeflation hollow of the Slowinski National Park, Poland. Acta Soc. Bot. Pol. 3,223–236.

Borstler, B., Renker, C., Kahmen, A., Buscot, F., 2006. Species composition ofarbuscular mycorrhizal fungi in two mountain meadows with different man-agement types and levels of plant biodiversity. Biol. Fert. Soils 42, 286–298.

Brundrett, M., Abbott, L.K., Jasper, D.A., 1999. Glomalean mycorrhizal fungi fromtropical Australia. I. Comparison of the effectiveness and specificity of differentisolation procedures. Mycorrhiza 8, 305–314.

Brundrett, M., Melville, L., Peterson, L., 1994. Practical Methods in MycorrhizaResearch. Mycologue Publications, University of Guelph, Guelph, Ontario,Canada.

Burrows, R.L., Pfleger, F.L., 2002. Host responses to AMF from plots differing in plantdiversity. Plant Soil 240, 169–179.

Clapp, J.P., Young, J.P.W., Merryweather, J.W., Fitter, A.H., 1995. Diversity of fungalsymbionts in arbuscular mycorrhizae from a natural community. New Phytol.87, 259–265.

Cuenca, G., De Andrade, Z., Escalante, G., 1998. Arbuscular mycorrhizae in therehabilitation of fragile degraded tropical lands. Biol. Fert. Soils 26, 107–111.

Daniell, T.J., Husband, R., Fitter, A.H., Young, J.P.W., 2001. Molecular diversity ofarbuscular mycorrhizal fungi colonising arable crops. FEMS Microbiol. Ecol. 36,203–209.

Douds, D.D., Millner, P., 1999. Biodiversity of arbuscular mycorrhizal fungi inagroecosystems. Agric. Ecosyst. Environ. 74, 77–93.

FAO, 2006. World reference base for soil resources. World Soil Resources Reports,vol. 103. Food and Agricultural Organisation of the United Nations (FAO) &International Soil Reference and Information Centre (ISRIC) & InternationalSociety of Soil Science (ISSS), Rome, Italy, p. 88.

Gai, J.P., Christie, P., Feng, G., Li, X.L., 2006. Twenty years of research on biodiversityand distribution of arbuscular mycorrhizal fungi in China: a review. Mycorrhiza16, 229–239.

Gange, A.C., 1993. Translocation of mycorrhizal fungi by earthworms during earlysuccession. Soil Biol. Biochem. 25, 1021–1026.

Gavito, M.E., Varela, L., 1993. Seasonal dynamics of mycorrhizal associations inmaize fields under low input agriculture. Agric. Ecosyst. Environ. 45, 275–282.

Gazey, C., Abbott, K.K., Robson, A.D., 1992. The rate and development of mycorrhizasaffects the onset of sporulation and production of external hyphae of twospecies of Acaulospora. Mycol. Res. 96, 643–650.

Gazey, C., Abbott, K.K., Robson, A.D., 1993. VA mycorrhizal spores from 3 species ofAcaulospora—germination, longevity and hyphal growth. Mycol. Res. 97, 785–790.

Gemma, J.N., Koske, R.E., 1988. Seasonal variation in spore abundance and dor-mancy of Gigaspora gigantea and in mycorrhizal inoculum potential of a dunesoil. Mycologia 80, 211–216.

Giovannetti, M., 1985. Seasonal variation of vesicular arbuscular mycorrhizas andEndogonales spores in a maritime sand dune. Trans. Br. Mycol. Soc. 84, 679–684.

Goto, B.T., Maia, L.C., 2006. Glomerospores: a new denomination for the spores ofGlomeromycota, a group molecularly distinct from the Zygomycota. Mycotaxon96, 129–132.

Hart, M.M., Reader, R.J., Klironomos, J.N., 2001. Life strategies of arbuscular mycor-rhizal fungi in relation to their successional dynamics. Mycologia 93, 1186–1194.

Hart, M.M., Reader, R.J., Klironomos, J.N., 2003. Plant coexistence mediated byarbuscular mycorrhizal fungi. Trends Ecol. Evol. 18, 418–423.

Hayman, D.S., 1970. Endogone spore numbers in soil and vesicular-arbuscularmycorrhiza in wheat as influenced by season and soil treatment. Trans. Br.Mycol. Soc. 54, 53–63.

Helgason, T., Daniell, T.J., Husband, R., Fitter, A.H., Young, J.P.W., 1998. Ploughing upthe wood-wide web? Nature 394, 431.

Hendrix, J.W., Guo, B.Z., An, Z.Q., 1995. Divergence of mycorrhizal communities incrop production systems. Plant Soil 170, 131–140.

Hijri, I., Sykorova, Z., Oehl, F., Ineichen, K., Mader, P., Wiemken, A., Redecker, D.,2006. Communities of arbuscular mycorrhizal fungi in arable lands are notnecessarily low in diversity. Mol. Ecol. 15, 2277–2289.

Husband, R., Herre, E.A., Young, J.P.W., 2002. Temporal variation in the arbuscularmycorrhizal communities colonising seedlings in a tropical forest. FEMS Micro-biol. Ecol. 42, 131–136.

F. Oehl et al. / Agriculture, Ecosystems and Environment 134 (2009) 257–268268

Jansa, J., Mozafar, A., Anken, T., Ruh, R., Sanders, I.R., Frossard, E., 2002. Diversity andstructure of AMF communities as affected by tillage in a temperate soil.Mycorrhiza 12, 225–234.

Johnson, N.C., Tilman, D., Wedin, D., 1992. Plant and soil controls on mycorrhizalcommunities. Ecology 73, 2034–2042.

Klironomos, J.N., McCune, J., Hart, M., Neville, J., 2000. The influence of arbuscularmycorrhizae on the relationship between plant diversity and productivity. Ecol.Lett. 3, 137–141.

Klironomos, J.N., Hart, M., 2002. Colonization of roots by arbuscular mycorrhizalfungi using different sources of inoculum. Mycorrhiza 12, 181–184.

Landis, F.C., Gargas, A., Givnish, T.J., 2004. Relationship among arbuscular mycor-rhizal fungi, vascular plants and environmental conditions in oak savannas.New Phytol. 164, 493–504.

Lee, P.J., Koske, R.E., 1994. Gigaspora gigantea: seasonal abundance and ageing ofspores in a sand dune. Mycol. Res. 98, 453–457.

Liu, Y., He, L., An, L.Z., Helgason, T., Feng, H.Y., 2009. Arbuscular mycorrhizaldynamics in a chronosequence of Caragana korshiniskii plantations. FEMSMicrobiol. Ecol. 67, 81–92.

Mader, P., Fliessbach, A., Dubois, D., Gunst, L., Fried, P., Niggli, U., 2002. Soil fertilityand biodiversity in organic farming. Science 296, 1694–1697.

Mandyam, K., Jumpponen, A., 2008. Seasonal and temporal dynamics of arbuscularmycorrhizal and dark septate endophytic fungi in a tallgrass prairie ecosystemare minimally affected by nitrogen enrichment. Mycorrhiza 18, 145–155.

Mathimaran, N., Ruh, R., Vullioud, P., Frossard, E., Jansa, J., 2005. Glomus intraradicesdominates arbuscular mycorrhizal communities in a heavy textured agricul-tural soil. Mycorrhiza 16, 61–66.

Merrywheather, J., Fitter, A.H., 1998a. The arbuscular mycorrhizal fungi of Hya-cinthoides non-scripta I. Diversity of fungal taxa. New Phytol. 138, 117–129.

Merrywheather, J., Fitter, A.H., 1998b. The arbuscular mycorrhizal fungi of Hya-cinthoides non-scripta II. Seasonal and spatial patterns of fungal populations.New Phytol. 138, 131–142.

Miller, D.D., Domoto, P.A., Walker, C., 1985. Mycorrhizal fungi at 18 apple rootstockplantings in the United States. New Phytol. 100, 379–391.

Mummey, D.L., Rillig, M.C., Holben, W.E., 2005. Neighbouring plant influences onarbuscular mycorrhizal fungal community composition as assessed by T-RFLPanalysis. Plant Soil 271, 83–90.

Oehl, F., de Souza, F.A., Sieverding, E., 2008. Revision of Scutellospora and descriptionof five new genera and three new families in the arbuscular mycorrhiza-forming Glomeromycetes. Mycotaxon 106, 311–360.

Oehl, F., Redecker, D., Sieverding, E., 2005a. Glomus badium, a new sporocarpicarbuscular mycorrhizal fungal species from European grasslands with highersoil pH. J. Appl. Bot. Food Qual. – Angew. Bot. 79, 38–43.

Oehl, F., Sieverding, E., Ineichen, K., Mader, P., Boller, T., Wiemken, A., 2003a. Impactof land use intensity on the species diversity of arbuscular mycorrhizal fungi inagroecosystems of Central Europe. Appl. Environ. Microbiol. 69, 2816–2824.

Oehl, F., Sieverding, E., Ineichen, K., Ris, E.A., Boller, T., Wiemken, A., 2005b. Commu-nity structure of arbuscular mycorrhizal fungi at different soil depths in exten-sively and intensively managed agroecosystems. New Phytol. 165, 273–283.

Oehl, F., Sieverding, E., Mader, P., Dubois, D., Ineichen, K., Boller, T., Wiemken, A.,2004. Impact of long-term conventional and organic farming on the diversity ofarbuscular mycorrhizal fungi. Oecologia 138, 574–583.

Oehl, F., Wiemken, A., Sieverding, E., 2003b. Glomus aureum, a new sporocarpicarbuscular mycorrhizal fungal species from European grasslands. J. Appl. Bot. –Angew. Bot. 77, 111–115.

Oehl, F., Sykorova, Z., Redecker, D., Wiemken, A., Sieverding, E., 2006. Acaulosporaalpina, a new arbuscular mycorrhizal fungal species characteristic for highmountainous and alpine regions of the Swiss Alps. Mycologia 98, 286–294.

Opik, M., Moora, M., Liira, J., Zobel, M., 2006. Composition of root-colonizingarbuscular mycorrhizal fungal communities in different ecosystems aroundthe globe. J. Ecol. 94, 778–790.

Pietrowski, J.S., Rillig, M.C., 2008. Succession of arbuscular mycorrhizal fungi: pat-terns, causes and considerations for organic agriculture. Adv. Agron. 97, 111–130.

Pietrowski, J.S., Lekberg, Y., Herner, M.J., Ramsey, P.W., Rillig, M.C., 2008. Dynamicsof mycorrhizae during development of riparian forests along an unregulatedriver. Ecography 31, 245–253.

Pringle, A., Bever, J.D., 2002. Divergent phenologies may facilitate the coexistence ofarbuscular mycorrhizal fungi in a North Carolina grassland. Am. J. Bot. 89, 1439–1446.

Rabadin, S.C., Stinner, B.R., 1989. The significance of vesicular-arbuscular mycor-rhizal fungal-soil interactions in agroecosystems. Agric. Ecosyst. Environ. 27,195–204.

Redecker, D., 2000. Specific PCR primers to identify arbuscular mycorrhizal fungiwithin colonized roots. Mycorrhiza 10, 73–80.

Redell, P., Spain, A.V., 1991. Earthworms as vectors of viable propagules of mycor-rhizal fungi. Soil Biol. Biochem. 23, 767–774.

Renker, C., Weisshuhn, K., Kellner, H., Buscot, F., 2006. Rationalizing molecularanalysis of field-collected roots for assessing diversity of arbuscular mycor-rhizal fungi: to pool, or not to pool, that is the question. Mycorrhiza 16, 525–531.

Richter, B.S., Stutz, J.C., 2002. Mycorrhizal inoculation of Big Sacaton: implicationsfor grassland restoration of abandoned agricultural fields. Rest. Ecol. 10, 607–616.

Rodriguez, A., Clapp, J.P., Dodd, J.C., 2004. Ribosomal RNA gene sequence diversity inarbuscular mycorrhizal fungi (Glomeromycota). J. Ecol. 92, 986–989.

Santos, J.C., Finlay, R.D., Tehler, A., 2006. Molecular analysis of arbuscular mycor-rhizal fungi colonising a semi-natural grassland along a fertilisation gradient.New Phytol. 172, 159–168.

Schenck, N.C., Perez, Y., 1990. Manual for the Identification of VA Mycorrhizal Fungi,3rd ed. Synergistic Publications, Gainesville, FL, USA.

Sieverding, E., Oehl, F., 2006. Revision of Entrophospora and description of Kuklos-pora and Intraspora, two new genera in the arbuscular mycorrhizal Glomer-omycetes. J. Appl. Bot. Food Qual. – Angew. Bot. 80, 69–81.

Smith, M.R., Charvat, I., Jacobson, R.L., 1998. Arbuscular mycorrhizae promoteestablishment of prairie species in a tallgrass prairie restoration. Can. J. Bot.76, 1947–1954.

Smith, S.E., Read, D.J., 2008. Mycorrhizal Symbiosis, 3rd ed. Academic Press, London.Spain, J.L., Sieverding, E., Oehl, F., 2006. Appendicispora: a new genus in the

arbuscular mycorrhiza-forming Glomeromycetes, with a discussion of thegenus Archaeospora. Mycotaxon 97, 163–182.

Stukenbrock, E.H., Rosendahl, S., 2005. Distribution of dominant arbuscular mycor-rhizal fungi among plant species in undisturbed vegetation of a coastal grass-land. Mycorrhiza 15, 497–503.

Sturmer, S.L., Bellei, M.M., 1994. Composition and seasonal variation of sporepopulations of arbuscular mycorrhizal fungi in dune soils on the island ofSanta Catarina, Brazil. Can. J. Bot. 72, 359–363.

Sylvia, D.M., 1986. Spatial and temporal distribution of vesicular arbuscular mycor-rhizal fungi associated with Uniola paniculata in Florida foredunes. Mycologia78, 353–358.

Titus, J.H., Whitcomb, S., Pitoniak, H.J., 2007. Distribution of arbuscular mycorrhizaein relation to microsites on primary successional substrates on Mount St.Helens. Can. J. Bot. 85, 941–948.

van der Heijden, M.G.A., Boller, T., Wiemken, A., Sanders, I.R., 1998a. Differentarbuscular mycorrhizal fungal species are potential determinants of plantcommunity structure. Ecology 79, 2082–2091.

van der Heijden, M.G.A., Klironomos, J.N., Ursic, M., Moutoglis, P., Streitwolf-Engel,R., Boller, T., Wiemken, A., Sanders, I.R., 1998b. Mycorrhizal fungal diversitydetermines plant biodiversity, ecosystem variability and productivity. Nature396, 69–72.

Walker, C., 2008. Ambispora and Ambisporaceae resurrected. Mycol. Res. 112, 297–298.

Walker, C., Vestberg, M., Demircik, F., Stockinger, H., Saito, M., Sawaki, H., Nishmura,I., Schussler, A., 2007. Molecular phylogeny and new taxa in the Archaeosporales(Glomeromycota): Ambispora fennica gen. sp. nov., Ambisporaceae fam. nov., andemendation of Archaeospora and Archaeosporaceae. Mycol. Res. 111, 137–153.

Wu, B.Y., Hogetsu, T., Isobe, K., Ishii, R., 2007. Community structure of arbuscularmycorrhizal fungi in a primary successional volcanic desert on the southeastslope of Mount Fuji. Mycorrhiza 17, 495–506.

Wu, T.H., Hao, W.Y., Lin, X.G., Shi, Y.Q., 2002. Screening of arbuscular mycorrhizalfungi for the revegetation of eroded red soils in subtropical China. Plant Soil 239,225–235.