Community structure of arbuscular mycorrhizal fungi at different soil depths in extensively and intensively managed agroecosystems


Author for correspondence: Andres Wiemken Tel: +41 (0)61 2672310 Fax: +41 (0)61 2672330 Email:


  • • The vertical distribution of spores of arbuscular mycorrhizal fungi (AMF) was investigated in soil profiles of extensively and intensively managed agroecosystems, including two permanent grasslands, a vineyard and two continuously mono-cropped maize fields.
  • • The number of AMF spores decreased with increasing soil depth – most drastically in the grasslands and the vineyard – but there was a large diversity of AMF species even in the deepest soil layers (50–70 cm). This was particularly striking in the maize fields where the highest species numbers were found below ploughing depth. Some species sporulated mainly, or exclusively, in the deep soil layers, others mainly in the top layers.
  • • Soil samples were used to inoculate trap cultures. Up to 18 months after inoculation, there was no conspicuous difference in the species composition among the trap cultures representing different soil depths, and only a weak match to the species composition determined by analysis of field samples.
  • • Our results indicate that the AMF communities in deep soil layers are surprisingly diverse and different from the topsoil. Thus, deep soil layers should be included in studies to get a complete picture of AMF diversity.


Ecological studies on the community structure of arbuscular mycorrhizal fungi (AMF) are generally restricted to the main rooting zone (10–30 cm soil depth, e.g. Stutz & Morton, 1995, Douds et al., 1995, Guadarrama & Alvarez-Sanchez, 1999; Bever et al., 2001) regardless of whether they were based on spore morphotyping (e.g. Blaszkowski, 1994; Kurle & Pfleger, 1996; Franke-Snyder et al., 2001) or on molecular techniques allowing identification of AMF species directly in the roots (Helgason et al., 1998, 2002; Husband et al., 2002; Redecker, 2002; Vandenkoornhuyse et al., 2002; Jansa et al., 2003; Johnson et al., 2003). Only a few studies included the subsoil. With increasing soil depth, a decrease was found in the percentage of roots colonized by AMF (Jakobsen & Nielsen, 1983; Rillig & Field, 2003), in the number of infective propagules (An et al., 1990), in the amount of extraradical AMF hyphae (Kabir et al., 1998, already in 15–25 cm soil depth when compared to the upper 15 cm), in AMF spore (Jakobsen & Nielsen, 1983; Zajicek et al., 1986; Thompson, 1991) and in species numbers (Zajicek et al., 1986). The latter authors detected only two AMF species deeper than 40 cm and only one species deeper than 60 cm in a prairie grassland of the Great Plains. Nothing has been reported about the vertical distribution of AMF species in soils under extensive compared with intensive agricultural management. This is of interest because soil management methods and agronomic practices may affect the AMF community structure positively or negatively, to the benefit or disadvantage of crop yields and land productivity (van der Heijden et al., 1998; Mäder et al., 2002).

In the present study, we investigate the abundance and diversity of AMF spores at different soil depths, considering three agricultural land use systems widely differing in management intensity and prevailing in the Upper Rhine Valley: extensively managed, permanent grasslands, a vineyard managed at intermediate intensity, and intensively managed, continuously mono-cropped maize fields. We have previously analysed the abundance and diversity of AMF spores in topsoils of both the grasslands and the maize fields, using samples from the same field sites, and this previous study has clearly demonstrated that agricultural intensification, as practiced in temperate Central Europe, negatively affects AMF abundance and diversity in the topsoils (Oehl et al., 2003a). We now show that deeper soil layers differ considerably in the abundance and diversity of AMF species, as analysed by morphotyping of spores, and that in intensively managed agroecosystems, these deep soil layers may represent a hidden source of additional AMF diversity. This insight is particularly relevant for current attempts to restore degraded soils that – due to detrimental agricultural or other landuse practices – have been impoverished with respect to AMF diversity (An et al., 1990; Cuenca et al., 1998a,b).

The present study is based on spore counts and on identification of the AMF by spore morphology. Since these AMF spores accumulated over a certain time (weeks to months), our study provides an integrative holistic, but somewhat static picture of the AMF community. Clearly, spore populations do not directly reflect the AMF community that is actually colonizing the plant roots (Clapp et al., 1995). In fact, some AMF may not be detected at all because they are not sporulating or are only doing so occasionally. In the future, it may be possible to obtain a more dynamic and more complete picture of the AMF community actually present and active in the roots, using molecular identification tools. However, this approach is currently limited due to the considerable costs involved, the lack of adequate primers to cover the whole range of AMF, and the difficulties to assign sequences to taxonomic units. For the time being, therefore, morphological identification of AMF spores remains the most economical, the fastest, and, indeed, the only feasible way to assess AMF communities in studies on soil samples taken from the field (Douds & Millner, 1999; Oehl et al., 2003a).

Materials and Methods

Sites and their agricultural practice

The Upper Rhine Valley between Basel (CH), Freiburg i. Br. (D) and Mulhouse (F), has 650–850 mm yearly precipitation and c. 9.5°C annual average temperature; its agricultural soils have mostly developed on periglacial Loess. We selected five field sites in this area, representing three different agricultural land use systems that differ in management intensity: two extensive grasslands, one vineyard and two intensively managed, continuously mono-cropped maize fields. The soil properties at these field sites are summarized in Table 1. One of the grasslands (G) and one of the maize fields (F) are de-carbonated (representing a Luvisol and an Alisol soil, respectively). The other grassland (V) and the other maize field (R) are situated on the hilly borders of the valley and therefore subjected to continuous slight surface erosion. They are still carbonated (representing Calcaric Regosols), and they have low clay contents, which decrease further with increasing soil depth.

Table 1.  Chemical and physical soil properties in different soil depths at the field sites and in the trap culture substrate
 Site abbreviation (soil type)Soil depth (cm)pH (KCl)Organic Carbon (g C kg−1)Phosphorus (Na-acetate)a (mg P kg−1)Phosphorus (Double lactate)a (mg P kg−1)Clay content (%)
  • a

    Soil extractant; see the Material and Methods section.

  • b

    The high clay contents in the topsoil of grassland G when compared to site F suggest that this decarbonated Loess soil had been subjected to significant soil erosion during the last centuries, after previous de-carbonation and clay mineral formation.

  • c

    On the soil type level Luvisols and Alisols are differentiated through the percentage of base saturation (pH; FAO, 1998).We assume that the high-intensity use of acidifying mineral fertilizers at site F is responsible for the development of the soil from a Luvisol to an Alisol.

  • d

    The vineyard is located in the direct neighbourhood of grassland V and has been rigoled, i.e. the soil has been turned over to 80 cm depth before the young grapes were planted c. 20 year ago. Through this widespread viticultural practice, the soil type changed from a Regosol to an Anthrosol according to FAO (1998).

GrasslandsV (Calcaric Regosol) 0–107.044.7 5.210.916
10–207.623.2 8.3 8.714
20–357.7 9.3 7.910.511
35–507.8 4.1 6.5 7.9 9
50–707.9 2.3 7.0 7.4 9
G (Haplic Luvisol)b,c 0–106.437.1 7.421.830
35–506.611.012.2 9.234
50–707.0 5.811.4 7.933
VineyardJ (Regic Anthrosol)d 0–107.413.317.025.813
10–207.8 5.811.8 9.612
20–357.9 3.5 6.5 7.012
35–507.9 3.5 6.1 6.511
50–708.0 2.9 6.1 6.510
Maize fieldsR (Calcaric Regosol) 0–107.010.410.518.317
20–357.8 4.6 7.4 8.712
35–507.8 2.9 7.0 8.310
50–707.9 2.3 7.0 7.9 7
F (Haplic Alisol)c 0–104.410.410.944.116
35–504.0 4.615.721.823
50–704.0 3.513.520.527
Trap culture substrate  6.2 0.5 3.029.7< 5

All sites, except the vineyard, have already been investigated in a previous study, which reported the AMF spore communities in the topsoils; details about land use, geographical position, and numbers of grassland plant species can be found there (Oehl et al., 2003a). The two maize fields selected differ in one major respect: Tillage at site R is more shallow (20 cm) than at site F (35 cm). The vineyard is located in the direct neighbourhood of grassland V. It regularly receives 60–80 kg N and 30–40 kg P ha−1 a−1, and fungicides and insecticides are applied according to the guidelines of German Viticulture, respecting threshold dressings. The interrows between the grapes are subject to a 2-yr cycle of treatments (alternating from interrow to interrow), namely first cultivation by chiseling twice per year and second leaving the soil un-cultivated but mowing the up-coming vegetation (c. 15 species of typical vineyard plants) twice per year. At the five field sites, the respective agricultural land use system had not been changed for the last 30–40 yr, except that at the sites F and R winter wheat was grown instead of maize in some years, but not during the last 5–8 yr previous to the sampling.

Soil sampling, preparation and chemical soil analysis

Soil samples were taken at each field site in four replicate plots of 100 m2 during October 2000. During this month, the growing season of grasslands, grapes and maize ends in the study area. From the grasslands and maize fields, topsoil had been sampled already in spring 2000 (from same replicate plots), that is at the beginning of the growing season; the results regarding AMF spore diversity have already been reported (Oehl et al., 2003a). In each of the four replicate plots per field site, six soil cores were taken to a depth of 70 cm using a Pürckhauer type corer (length 100 cm; Agrolab-Gruppe, Langenbach-Oberhummel, Germany), and the cores were separated in sections corresponding to 0–10, 10–20, 20–35, 35–50 and 50–70 cm soil depth. In the vineyard, three cores from the cultivated rows and three cores from the noncultivated interrows were taken in each replicate plot. The six individual core segments of a given plot that corresponded to a certain soil depth were pooled and analysed as a single sample. The samples were carefully ground by hand, thoroughly mixed, and air-dried. They were kept at 4°C until analysing AMF spore abundance and species richness. Selected chemical and physical soil parameters (pH, organic carbon, available phosphorus, clay content) for each soil layer are presented in Table 1. They were measured in the laboratory of F.M. Balzer, Wetter-Amönau, Germany (, according to standard methods; phosphorous (P) is extracted with sodium acetate (‘P-Morgan’) and double lactate (‘P-DL’) (methods for estimating plant-available P throughout Europe: Neyroud & Lischer, 2003).

Trap cultures

For trap culture analysis, the two topsoil samples (0–10 and 10–20 cm) as well as the subsoil samples (20–35, 35–50 and 50–70 cm) were pooled. Only for site F, where deep tillage (35 cm) was applied, samples from 0 to 35 cm and from 35 to 70 cm were pooled.

Trap cultures were established and managed as described previously (Oehl et al., 2003a). Briefly, one container (300 × 200 × 200 mm) per field plot replicate and soil depth-group was equipped with a 20 mm thick drainage mat and filled with 3 kg of autoclaved substrate (TerraGreen-Loess mixture; 3 : 1, w/w; the chemical and physical parameters of this substrate are shown in Table 1). Three plantlets each of Lolium perenne, Trifolium pratense and Plantago lanceolata were used as host plants in each trap culture. The AMF inoculum consisted of nine samples of largely undisturbed soil core pieces (in total 180 g d. wt per container), which were placed directly below the spots where the plantlets were inserted later on. Four replicate containers were set up also with a nonmycorrhizal control treatment, allowing to check the risk of contamination. The trap cultures were kept in a glasshouse for 18 months under natural ambient light and temperature conditions until July 2002. During this period, the trap plants were cut repeatedly 3 cm above the ground. From each container, four cores (15 cm3, sampling depth 10 cm) were taken after 6, 12 and 18 months for AMF spore identification. No AMF sporulation was detected in the nonmycorrhizal controls over 18 months.

AMF spore isolation and identification

The AMF spores occurring in the original field samples, or produced in the trap cultures, were extracted from the soil/substrate by wet sieving and sucrose density gradient centrifugation as described before (Oehl et al., 2003a). The procedure included passing 25 g of air-dried field samples, or 60 cm3 of harvested trap culture substrate through a set of sieves with 1000, 500, 125 and 32 µm mesh width. Spores were counted using a dissecting microscope at up to 90 fold magnification. About 40–70% of the spores were mounted on slides in polyvinyl–lactic acid–glycerine (PLVG) (Koske & Tessier, 1983) or in a mixture (1 : 1; v/v) of PLVG with Melzer's reagent (Brundrett et al., 1994). Only the healthy looking spores were mounted and subsequently examined using a compound microscope at up to 400-fold magnification.

In this study, the term ‘species’ either stands for a known and clearly identified species based on spore morphology, or for a not yet known, newly identified species, or for a putative identified known species that is not totally clear to the authors, or for a ‘species group’ comprising more than one identified species that generally cannot be properly discerned (Oehl et al., 2003a). Such a ‘species group’ was counted as one species in our analysis of the AMF species number per site. However, for the summarizing final presentation of the total number of species detected in the whole study, all species clearly identified at least once were counted. As far as possible, identification was based on current species descriptions and identification manuals (Schenck & Peréz, 1990; International Culture Collection of Arbuscular and Vesicular-Arbuscular Endomycorrhizal Fungi:

Spore abundance at field sites is presented as total number of spores extracted per gram soil. The number of AMF species detected per g soil, that is the species richness of a site, depends on the number of individuals present in a sample. Hence, the species richness at the field sites was corrected using the rarefaction method (Legendre & Legendre, 1998).

The Shannon-Weaver (H′) index was calculated as an additional measure of AMF diversity, as it combines two components of diversity, species richness and evenness. It is calculated from the equation H′ = -Σρi ln ρi, where ρi is the relative spore abundance of the ith species compared with all species identified in a sample.


Significance of differences in spore abundance, species richness and diversity (Shannon-Weaver index) between the samples from different soil depths at a given site was tested using Fisher's least significant difference (LSD) at P < 0.05 after a one-way analysis of variance (anova).


AMF spore abundance

The spore abundance was highest in the topsoils regardless of the land use system, and decreased with soil depth in all soils investigated (Fig. 1). This decrease was most pronounced in the grassland soils where 69–84 spores g−1 were found in the topsoil and only 3.0–3.3 at a depth of 50–70 cm. At every soil depth, considerably more spores were found in the grasslands than in the maize fields, where only 7–14 spores g−1 were found in the topsoils, and only c. 1.3 spores g−1 in the layer at 50–70 cm depth. In terms of spore abundance, the vineyard spore numbers and vertical distribution were intermediate between the grasslands and the maize fields (Fig. 1).

Figure 1.

Spore abundance in different soil depths at the field sites subjected to different agricultural land use. Data are reported as averages (+ SD) for four replicate plots per site. Nonsignificant differences between soil layers of the same site 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.

AMF species

A total of 30 morphologically distinctive AMF species were found at the field sites (Table 2). Most of them belonged to the genus Glomus (G). Two species were from the genus Scutellospora C. Walker & F.E. Sanders (S. calospora and S. castanea), while two belonged to the genus Acaulospora Gerd. & Trappe (A. laevis and A. paulinae), one to Entrophospora R.N. Ames & R.W. Schneid. (E. infrequens) and one to Paraglomus J.B. Morton & D. Redecker (P. occultum). One of the species was recently placed from Glomus into the new genus Pacispora Sieverd. & Oehl (Pa. dominikii; Oehl & Sieverding, 2004). Three Glomus species could not be attributed to an AMF species described so far (Glomus sp. strains BR4, BR8 and BR9). Remarkably, the two Acaulospora species were recovered only from de-carbonated soils.

Figure 2.

Relative spore abundancea (in percentage) of arbuscular mycorrhizal fungi (AMF) species found in different soil depthsb at the field sites

Vertical distribution of AMF species

The vertical spore distribution differed considerably for different AMF species. For several species, the relative spore abundance in relation to the total AMF spore populations apparently did not change (Group A, Table 2). These AMF were, for instance, Glomus geosporum and G. constrictum. The proportion of another group of species was decreasing with depth in the majority of the sites where they were detected (Group B, Table 2). These were, for instance, G. fasciculatum and G. diaphanum. In particular, several sporocarpic Glomus spp. – only occurring in the grassland soils – showed a decreasing relative spore abundance with increasing soil depth (G. sp. strain BR8 and G. microcarpum) or were not at all found in the deeper soil layers investigated (G. rubiforme and G. sinuosum). Spores of other species were increasingly found with increasing soil depth, at least in relative terms (Group C, Table 2). These were, for example, G. etunicatum and G. aureum and especially the Scutellospora species detected (S. calospora and S. castanea). In the maize fields, several other species of this group were exclusively found in the subsoils. These were E. infrequens and A. paulinae (site F), G. invermaium (site R) and Pacispora dominikii (site R and F). A few other AMF showed varying vertical distribution patterns throughout the soil profile in different land use systems (Group D, Table 2). When spores of G. caledonium and in particular of G. mosseae occurred at a site, they usually showed a decreasing absolute and relative spore abundance in the soil profile of the maize fields and the vineyard, but were – at least in relative terms – increasingly found in the lower soil depths of the grasslands (Table 2). The relative abundance of G. mosseae spores increased in the grasslands from 0.5 to 2% in the topsoils to 4–5% in the deeper layers and, conversely, decreased in the vineyard and maize fields from 12 to 58% in the topsoils to 4–7% in the subsoil. At the grassland sites, spores of – usually sporocarpic –G. macrocarpum decreased similar to those of G. rubiforme and G. sinuosum with increasing soil depth, but, remarkably, increased in the vineyard. This species, however, was absent in the high-input maize fields.

AMF species richness

The AMF species richness for each soil depth at the sites is presented as the average of the species numbers found in the samples from the replicate plots and, in addition, as the total number of species detected per site at each soil depth (Fig. 2). In the upper three layers (0–10, 10–20 and 20–35 cm), these numbers were highest in the grasslands, slightly lower in the vineyard and lowest (c. 60% of that in the grasslands) in the maize fields (Fig. 2). The species richness decreased in the grasslands and in the vineyard continuously with increasing soil depth. By contrast, in the maize fields the highest richness was found below ploughing depth. This finding was more pronounced in field R where shallow tillage practice has been applied (20 cm) than in field F subjected to deeper (35 cm) tillage (Fig. 2). The differences in species richness between the grasslands and the maize fields were minor in the deepest soil layer investigated (50–70 cm). There, the richness was lowest in the deeply tilled maize field F and in the vineyard.

Figure 2.

Numbers of arbuscular mycorrhizal fungi (AMF) species in different soil depths at the field sites subjected to different agricultural land use. Data are reported as averages (+ SD) for four replicate plots per site and soil layer, and (solid diamond) as total numbers per soil layer (sum of four replicate plots). Nonsignificant differences between soil layers of the same site 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.

AMF diversity (Shannon-Weaver index)

In the two upper soil layers of the grasslands and the vineyard, the values of the Shannon-Weaver diversity index were higher (1.9–2.1) than in the maize fields (1.4–1.8) (Fig. 3). In the grasslands, the vineyard and the maize field F with deep tillage, the values decreased with increasing soil depth. By contrast, in the maize field R with shallow tillage, highest AMF diversity was found below the tilled layers (in 20–35 and 35–50 cm soil depth; Fig. 3) where the diversity index reached the values of the grassland topsoils, despite the lower species richness.

Figure 3.

Arbuscular mycorrhizal fungi (AMF) diversity (Shannon-Weaver index) in different soil depths at the field sites subjected to different agricultural land use. Diversities are reported as averages (+ SD) for four replicate plots per site. Nonsignificant differences between soil layers of the same site 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.

AMF species in the trap cultures

Twenty-five AMF species produced spores in the trap cultures within 18 months (Table 3). From the trap cultures inoculated with the grassland soils V and G, 18 and 20 AMF species were recovered, respectively, while only 11 were recovered from the vineyard and 14 and 13 in the maize fields R and F, respectively.

Figure 3.

Time perioda after trap culture initiation at which spores of different arbuscular mycorrhizal fungi (AMF) species were detected for the first time

There was a considerable discrepancy between the results of a direct analysis of soil samples and of the corresponding trap cultures. On the one hand, eight of the AMF detected in the trap cultures had not been observed in any of the field samples (Group E, Table 3). Furthermore, G. aggregatum, spores of which could not be detected in any field samples of grasslands, formed spores in all four trap cultures from the grasslands. On the other hand, nine species identified in the samples from the field did not sporulate in any of the trap cultures. These were mainly sporocarpic AMF that had been detected only in the upper soil layers of the grasslands rich in organic matter (e.g. G. sinuosum and G. rubiforme) and, furthermore, AMF that occurred preferentially or exclusively at deep soil depths in the maize fields (e.g. S. castanea, E. infrequens and Pa. dominikii). As already observed in the field samples, none of the Acaulospora species detected in the de-carbonated grassland site G was recovered in the trap cultures derived from the carbonated sites V, J and R.

Species recovered from the trap cultures representing different soil depths

When comparing the two trap cultures representing topsoil and deeper soil layers (Table 3), some results fitted well to the corresponding results from direct analysis of field samples (Table 2). For example, spores of G. etunicatum and G. mosseae, which occurred in all soil depths when analysed directly, appeared within 6–12 months in almost all the trap cultures. G. fasciculatum was more abundant in the topsoils, and it generally appeared more rapidly, or exclusively, in the trap cultures representing the topsoils. Glomus aureum– at field site R only recovered from the subsoil (Table 2) – was again only detected in the trap cultures of the subsoil samples (Table 3). Glomus caledonium, more numerous in the subsoil than in the topsoil at grassland site V, was only recovered from the corresponding subsoil cultures. By trap culturing, additional information was gained for Glomus geosporum: It was recovered faster from the subsoil than from the topsoil cultures of the grasslands and the vineyard, but from the cultures of maize field R, it was only obtained from the topsoil. This finding suggests a similar vertical distribution pattern for G. geosporum as already found for G. mosseae and G. caledonium. However, the comparison also reveals discrepancies between direct spore analysis and analysis of spores formed in the trap cultures. For example, G. aureum clearly had a higher relative spore abundance in deep soil layers at both grassland sites, but was found only in the trap cultures from the topsoils of the grassland. Similarly G. invermaium, which was represented in all soil depths of grassland G, was found only in the trap cultures derived from the topsoils of that site. On the other hand, G. diaphanum, which was well represented in the samples from all soil depths in grassland G and maize field R, was only forming spores in the trap cultures derived from the corresponding deep soil layers but not in the ones from the top soils.


Previously, we have reported that intensification of land use (Oehl et al., 2003a) and conventional, as opposed to organic, farming practices (Oehl et al., 2004) cause a reduction in AMF spore abundance and AMF species diversity in the agroecosystems of Central Europe. According to common practice, we had analysed only the topsoil (0–10 cm depth) in these studies. In the present work, we investigated the soil over a large depth range. As expected, the topsoil layers from 0 to 20 cm depth contained by far the largest AMF spore abundances, a feature that was most expressed in the grasslands. More remarkably, however, the AMF community composition changed towards deeper soil layers and a surprisingly high species richness was observed even in the deepest soil layers examined (50–70 cm). Scutellospora species (S. calospora and S. castanea) in particular were found to occur more abundantly with increasing soil depth, in relative and sometimes even in absolute terms (Table 2). Scutellospora castanea was detected only below 20 cm soil depth. Thus, at least with respect to spore formation, the Scutellospora species appear to be specialized for deeper layers of the Loess soils in the agro-climatic region under study. The increase of Scutellospora spore abundance towards deeper soil layers was more pronounced in the intensively managed maize fields (particularly at site R) when compared with the less intensively managed vineyard and the extensive grasslands (Table 2). This observation agrees well with the recent finding in a long-term field trial – using molecular tools – that maize roots from plowed and chiseled plots were colonized by Scutellospora to a much lower extent than roots from no-till, less intensively managed plots (Jansa et al., 2003). In another long-term field trial comparing different farming systems (Mäder et al., 2002), the occurrence of S. calospora and S. pellucida spores (Oehl et al., 2004) was found to be negatively correlated with the soil contents of available phosphorus (Oehl et al., 2002). These findings suggest two possible reasons for the stimulation of development of S. calospora and S. castanea in deeper soil layers (Table 2), namely the reduced mechanical soil disturbance or the decreased supply of available phosphorus (Table 1).

In the present study, where sampling was carried out in autumn, a much lower percentage of spores could be positively identified than in a previous study where sampling took place in springtime at the same field sites, except in the vineyard (Oehl et al., 2003a). Only 15–50% of all spores isolated, or 30–70% of the spores mounted on slides could be assigned to a species or at least to a species group, compared with 30–60% of all isolated spores, or 60–85% of the spores mounted on slides in the previous study (Oehl et al., 2003a). Most likely, this is due to the problem that in autumn, a major proportion of the spores is still immature and cannot be identified. Nevertheless, a very similar AMF community composition was obtained for both the grasslands and the maize fields in these studies. Only for a few species, spore abundance significantly changed from springtime to autumn; Scutellospora calospora spores, for instance, were more abundant in spring than in autumn; in trap cultures under natural ambient light and temperature conditions, this species exhibited a burst of sporulation between October and December (Oehl et al., 2003a, 2004), which could explain our observation. The increased spore abundance of S. calospora towards the deep soil layers and the characteristic of late-seasonal sporulation of this species might be related since in the deep soil layers the soil temperature is more constant and, during autumn until early winter, generally higher than in the topsoil, which is exposed to the frequently low temperature during late fall in Central Europe.

Several species found frequently in the extensive grasslands were not found in the intensively managed maize fields, neither in the topsoils nor in the deeper soil layers (Tables 2 and 3). This finding is in agreement with the observation that the majority of these AMF (G. sp. strains BR8 and BR9, G. microcarpum, G. rubiforme and G. sinuosum) were strongly decreasing in abundance with increasing soil depth or were not detected at all in the subsoil layers of the grasslands (Table 2). It appears that these AMF, at least in Central Europe, preferentially inhabit undisturbed topsoil rich in organic matter as occurring in grasslands. Another possibility is that they might need specific plant hosts. Accordingly, most of them were not recovered in the trap cultures (Table 3) containing a substrate devoid of organic matter (Table 1). It is not clear why, in previous studies (Oehl et al., 2003a, 2004), some of these AMF species sporulated abundantly.

Glomus aggregatum, G. geosporum, G. constrictum, G. fasciculatum, G. diaphanum, G. tunicatum, G. caledonium and G. mosseae are commonly found even in intensively managed arable lands (Land & Schönbeck, 1991; Blaszkowski, 1993; Kurle & Pfleger, 1996; Franke-Snyder et al., 2001; Oehl et al., 2003a). These species are sometimes called ‘typical AMF of arable lands’ or AMF ‘generalists’ (Oehl et al., 2003a) or even AMF ‘weed’ species (JPW Young, pers. comm.). According to our study they appear to belong to different groups with respect to their vertical distribution in the soils (Groups A–D, Table 2); some of them (G. mosseae and G. caledonium) differed in this respect depending on the agroecosystem. We assume that even these AMF ‘generalists’ might fulfil different ecological functions.

AMF spore abundance and species richness in general decreased with increasing soil depth (Figs 1 and 2). The decreases were much more pronounced in the extensive grasslands and in the vineyard than in the intensively managed maize fields. In fact, in one instance (maize field R), the highest number of species (Fig. 2) and the highest diversity (Fig. 3) was found in the sample corresponding to a soil depth of 20–35 cm, just below ploughing depth. This finding suggests that several AMF species apparently eradicated from the intensively managed maize field (top)soils (Oehl et al., 2003a) may have found a refuge or at least a preferred habitat below ploughing depth and, thus, are not completely lost through the agricultural practices. These species were, for example Glomus invermaium, Pacispora dominikii, Acaulospora paulinae, Entrophospora infrequens and Scutellospora castanea (Table 2). In general, our finding corroborates observations in other studies indicating that intensification of tillage practices (Jansa et al., 2002) and high-input conventional farming, compared with low-input organic farming (Oehl et al., 2004), negatively affect AMF abundance and diversity, especially with respect to AMF species not belonging to the genus Glomus.

The possibility of the survival of sensitive AMF species in the subsoil under adverse conditions, caused by intensive farming practices deleterious to AMF diversity, has important implications. An AMF ‘gene bank’, may persist in the subsoil, facilitating agro-ecological restoration when switching from a high-input to a low-input farming system that has to rely more on internal biotic and abiotic resources.


We thank Rolf Spirig and Giacomo Busco for technical assistence in the glasshouse and Dorothy Mazlum, Thomas Ortega and Lucius Cueni for the practical work they performed during the ‘Winterpraktikum 2000’ in our laboratory. Dr Dirk Redecker is gratefully acknowledged for critically reading the manuscript. This study was supported by the Swiss Agency for Development and Cooperation (SDC) in the frame of the Indo-Swiss Collaboration in Biotechnology (ISCB) programme.