Phylogenetic divergence in a local population of the ectomycorrhizal fungus Cenococcum geophilum


  • G. W. Douhan,

    Corresponding author
    1. Department of Plant Pathology, One Shields Avenue, University of California at Davis, Davis, CA 956161, USA
      Author for correspondence: G. W. Douhan Tel: +1 530 754 9894Fax: +1 530 752 5674Email:
    Search for more papers by this author
  • D. M. Rizzo

    1. Department of Plant Pathology, One Shields Avenue, University of California at Davis, Davis, CA 956161, USA
    Search for more papers by this author

Author for correspondence: G. W. Douhan Tel: +1 530 754 9894Fax: +1 530 752 5674Email:


  • • Cenococcum geophilum is a widely distributed mycorrhizal species associated with diverse gymnosperm and angiosperm hosts. In previous studies, a significant amount of genetic and genotypic diversity has been detected in this species, despite the fact that C. geophilum is not thought to reproduce by meiotic or mitotic spores.
  • • We conducted a phylogenetic analysis of 103 C. geophilum isolates from a California oak woodland and seven non-California isolates using a glyceraldehyde 3-phosphate dehydrogenase gene. In addition, a subset of isolates was analyzed using sequences from ITS-rDNA, a Group I intron located in the 3′ end of the SSU-rDNA and a portion of the mitochondrial SSU-rDNA.
  • • Phylogenetically distinct lineages, or cryptic species, of C. geophilum were detected at the scale of a single soil sample within our field site. As much genetic diversity was found within a soil sample as was found for isolates collected across the USA.
  • • Our results help explain the large amount of physiological, phenotypic, and genetic differences reported among isolates of C. geophilum from similar as well as diverse geographic regions. The ecological role that these sympatric cryptic species play remains to be determined.


There are c. 100 000 described species of fungi (Kirk et al., 2001). However, it has been estimated that there are at least 1.4 million more species of fungi that have yet to be discovered that likely exist in understudied environments (Hawksworth & Rossman, 1997; Hawksworth, 2001). Compared with above-ground terrestrial environments, the soil supports a significant amount of biodiversity for many groups of organisms (Wardle, 2002) and probably harbors many undescribed species of fungi. A second source of new fungal species are cryptic species (i.e. species indistinguishable by typical morphological criteria or sexual compatibility). Phylogenetic analyses of DNA sequence data has made it possible to determine distinct evolutionary lineages that are the starting point for the recognition of distinct species from within previously described morphospecies (Taylor et al., 2000). Using phylogenetics to determine species in fungi is attractive because of the few morphological characters that many fungi have and because many are only known to reproduce asexually, making it not possible to use a biological species concept. Combining field studies with cultural and molecular genetic analyses has led to the ‘discovery’ of many new species in a wide range of fungal groups; it is a trend that will likely continue in the future as more environments are studied in greater detail (e.g. Hawksworth, 2001).

Cenococcum geophilum is one of the most commonly encountered soil fungi forming ectomycorrhizal (EM) associations with gymnosperms and angiosperms in diverse habitats throughout northern temperate regions (Trappe, 1964). C. geophilum is one of the few mycorrhizal species that is routinely identified based on the morphology of colonized roots. It can also be isolated directly from vegetative structures (sclerotia) found in soil and cultured in vitro. Given the wide host range, distribution, ease of experimental manipulation, and potential ecological importance as a mycorrhizal symbiont, a considerable amount of research has been conducted with this fungus regarding its physiology and ecology. Such studies have found considerable cultural and physiological variation among isolates collected from similar as well as diverse geographic regions (e.g. LoBuglio, 1999).

Since C. geophilum has not been shown to produce any types of spores or reproductive structures besides sclerotia, its taxonomic placement within the fungi remained unanswered until relatively recently. LoBuglio et al. (1996) provided the first evidence that C. geophilum is most closely related to a group of fungi known as the Loculoascomycetes (Phylum Ascomycota) based on an analysis of ribosomal DNA. This classification has also been recently supported by a phylogeny using RNA polymerase II (Liu & Hall, 2004). Despite the fact that C. geophilum is not thought to reproduce by meiotic or mitotic spores, recent population genetic analyses have revealed considerable genotypic diversity within and among populations of this fungus. Jany et al. (2002) detected 24 genotypes out of 42 isolates based on a combination of randomly amplified polymorphic DNA (RAPD) markers, restriction fragment length polymorphisms (RFLP) of the internal transcribed spacer (ITS) regions and 5.8S region of rDNA, and a sequence characterized amplified region (SCAR). Panaccione et al. (2001) identified 12 genotypes out of 13 isolates based on amplified fragment length polymorphisms (AFLP). LoBuglio and Taylor (2002) detected 19 and 11 genotypes out of 48 and 34 isolates, respectively, from two populations of C. geophilum based on single nucleotide polymorphic (SNP) markers.

Based on RFLP analysis of the entire rDNA region, it has been suggested that C. geophilum is either a very heterogeneous species or is a species complex (LoBuglio et al., 1991). LoBuglio et al. (1991) detected 32 genotypes out of 70 isolates collected from broad host and geographic ranges with some of this variation attributed to a Group-I intron (CgSSU intron) found within the 3′ end of the small subunit (SSU) of rDNA (LoBuglio, 1999). In a follow up study using the same isolates as LoBuglio et al. (1991), a phylogenetic analysis was conducted on the ITS-rDNA region, which revealed up to 4% sequence divergence among the isolates (Shinohara et al., 1999). Shinohara et al. (1999) concluded that C. geophilum was in fact a ‘single taxonomic entity, possibly a single species’ that was extremely adaptable and widespread. Their conclusion was partially based on the fact that other fungal species have been reported to have as much intraspecies diversity in the ITS regions as did C. geophilum (Shinohara et al., 1999). However, other fungi with very similar or identical ITS regions have also been considered to be different species based on ecological and physiological data (e.g. Harrington & Rizzo, 1999). Based on the above studies, it is still unclear if a species complex exists in C. geophilum, which is important to know when interpreting population genetic data.

Previous studies have concluded that a finer sampling scheme was needed to adequately describe the genetic structure of C. geophilum (Jany et al., 2002; LoBuglio & Taylor, 2002). Therefore, we chose to sample at a very fine spatial scale within our study site in an oak savannah woodland in Sierra Nevada foothills of California dominated by blue oak (Quercus douglasii). C. geophilum is commonly found on the roots of blue oaks and is thought to be an important mycorrhizal species in environments such as ours where water stress is prevalent (Mexal & Reid, 1973; Coleman et al., 1989). Our initial objective was to characterize the genetic structure of C. geophilum at our field site but preliminary analyses using AFLP suggested we were possibly dealing with a species complex based on distinctly divergent banding patterns from isolates collected from the same soil sample (G. W. Douhan, unpubl.). Since the presence of cryptic species within C. geophilum could bias estimates of population structure based on AFLP data, our first objective was to test the hypothesis that multiple phylogenetic lineages, and possibly cryptic species, of C. geophilum exist at a fine spatial scale.

Materials and Methods


Isolates of C. geophilum were collected from the Koch natural area within the University of California's Sierra Foothill Research & Extension Center located in Browns Valley, CA, USA, c. 100 km northeast of Sacramento (39°15′-N, 121°17′-W). The site consists of an annual grassland oak-woodland within the Sierra Nevada foothills that has been maintained as an undisturbed natural reserve since 1960. The climate is Mediterranean with hot, dry summers and mild, rainy winters with c. 70 cm annual mean precipitation and a mean annual temperature of 15°C (Dahlgren et al., 1997).

Three blue oaks were chosen within a closed canopy area of the reserve to sample three subpopulations of C. geophilum. Each mature oak tree was separated by 10–15 m and was considered to represent the location of each subpopulation for the purposes of this study. Oak seedlings distributed within a 3–4-m area around each tree (subpopulation) were used as sampling units. Eighteen to 20 seedlings per subpopulation were dug up using a shovel, equaling c. 1 l volume of soil, placed in plastic bags, and stored at 4°C until processing. Loose soil was carefully removed from the seedling roots and then the roots were gently washed under tap water to expose clusters of organic matter consisting of C. geophilum mycorrhizas, sclerotia, and soil. When clusters of mycorrhizas were not apparent, c. 50 ml by volume of soil was saved per sample to isolate C. geophilum sclerotia. Preliminary attempts at isolating C. geophilum from mycorrhizas yielded few cultures, therefore only sclerotia were used to obtain cultures for this study.

The organic clusters and/or soil samples were washed and filtered through two layers of cheesecloth and placed into Petri dishes containing deionized water. Sclerotia that floated in water were considered to be nonviable (J. M. Trappe, pers. comm.). Sclerotia that did not float were then surface sterilized in 30% hydrogen peroxide for c. 5–20 s and placed onto modified Melin-Norkrans's media (Marx, 1969) amended with 50 mg l−1 streptomycin (Sigma, St. Louis, MO, USA). The plates were incubated at room temperature in the dark for up to 1 month. All germinated sclerotia that grew were transferred to Malt Extract agar (Difco, Sparks, MD, USA) and incubated for an additional 1–3 months to allow enough growth for DNA extraction. The hyphae was then scraped off the agar plates, freeze dried, ground, and the DNA was isolated using a slightly modified phenol:chloroform extraction procedure of Lee & Taylor (1990). Additional isolates of C. geophilum from distant geographic origins (Maryland, Alaska, Oregon) were also included in the study for comparative purposes. Information regarding the isolates is given in Fig. 1 legend.

Figure 1.

Neighborjoining analysis of Cenococcum geophilum isolates based on glyceraldehyde 3-phosphate dehydrogenase gene (gpd). Isolates from the California oak-woodland are labeled using a three number code: first letter = subpopulation, second = sample location, and third = isolate number. Additional non-California isolates are A177 from Alaska (host: Betula nana), A145 from Oregon (host: Pseudotsuga menziesii), 03–4-II from Maryland oak savannah soil, and Btree1, I-1, I-2, & I-3 from Oregon (host: Quercus garryana). Three asterisks indicate isolates used in additional analyses (see Figs 2 and 3). Isolates labeled ‘no cgssu intron’ do not possess a Group I intron located in the 3′ end of the SSU-rDNA. Bootstrap support of 70% and above are indicated above nodes based on 1000 replicates. Sequences have been deposited in GenBank under accession numbers (AY818436AY818545).

Phylogenetic analyses

To screen for genetic diversity, the glyceraldehyde 3-phosphate dehydrogenase (gpd) gene was first amplified from all isolates (N = 110) using the primers gpd1 and gpd2 (Berbee et al., 1999). Preliminary analyses found this region to be informative and could be easily amplified from all isolates. The final 40 µl reaction mixture contained 2 µl of a 1 : 25 dilution of template DNA, 1X PCR buffer (Invitrogen, Carlsbad, CA, USA), 2.5 mm MgCl2, 0.2 mm each dNTP (Invitrogen), 3.75 µm of each primer, and 0.5 U of Taq polymerase (Invitrogen). Thermocycling conditions consisted of an initial hold at 94°C for 3 min, followed by 25 cycles of 94°C (30 s), 60°C (30 s), and 72°C (1 min), and a final hold of 72°C for 8 min. PCR products were cleaned using a Millipore Montage vacuum filtration system (Millipore Corp, Bedford, MA, USA) following the manufacturer's instructions and sequenced in one direction with gpd1 using a Big Dye® Terminator v3.1 Cycle Sequencing Kit (Applied Biosystems, Foster City, CA, USA). The reactions were then cleaned using Millipore's Montage SEQ96 vacuum filtration system (Millipore Corp, Bedford, MA, USA) following the manufacturer's instructions. Sequences were determined using an ABI 3730 XL capillary sequencer at the Genomic Facility (College of Agricultural & Environmental Sciences) at University of California at Davis.

The sequences were edited using Sequencher (version 4.1.2, Gene Codes Corporation, Ann Arbor, MI, USA) with care taken to only use sequences that produced strong signals since only one read was initially used. The sequences were aligned using Clustal X (version 1.81) (Thompson et al., 1997) and manually edited in MacClade version 4 (Maddison & Maddison, 2001). Neighbor Joining (NJ) analysis under the Kimura 2 parameter option and Maximum Parsimony (MP) analysis using the heuristic search procedure with 1000 random-addition-sequence replicates and tree-bisection-reconnection branch swapping were conducted using paup* version 4.0 beta 10 (Swofford, 2002). Confidence in tree topology was examined using bootstrap with 1000 replicates for NJ and 500 replicates for MP under the heuristic option. A maximum likelihood (ML) analysis was also done on a subset of isolates to determine if any algorithm bias existed in the data. The tree was midway rooted because there is no known close relative to C. geophilum (LoBuglio et al., 1996).

The gpd phylogeny was then used to choose isolates that represented the genetic diversity of C. geophilum and additional sequences were analyzed. The primers ITS1F and ITS4 (Gardes & Bruns, 1993) were used to amplify the ITS-rDNA regions and to determine if isolates possessed a Group I intron (CgSSU) located in the small SSU-rDNA. If an isolate did not have the CgSSU intron, then the amplified fragment was c. 500 bp shorter than isolates with the CgSSU intron (Shinohara et al., 1999). Isolates with the CgSSU intron were preferentially chosen to increase the amount of sequence data because most isolates contained the CgSSU intron (see Results) and this region has been shown to evolve at a rate that is c. 40% faster than the closely associated ITS region (Shinohara et al., 1999). Therefore, even though the ITS and CgSSU intron are tightly linked, they have the potential to yield different phylogenetic signals. We also found similar relationships among isolates without the CgSSU intron based on preliminary sequence analyses of the ITS and gpd regions so the selection of isolates without the CgSSU intron should not have introduced significant bias in the data. Sequencing reactions were set up as above using ITS1 and ITS4 (White et al., 1990) for the ITS-rDNA region and ITS1F and NS8 for the CgSSU intron (White et al., 1990; Gardes & Bruns, 1993) for a subset of isolates (n = 26). The second strand of the gpd gene was also sequenced from each of these representative isolates and all genes were edited and analyzed as above. A portion of the mitochondrial SSU ribosomal DNA (mit SSU rDNA) was also amplified and sequenced in both directions using the primers NMS1 and NMS2 (Li et al., 1994) from an additional subset of isolates (n = 12). Each data set was analyzed individually and the gpd, ITS, and CgSSU regions were combined following the ‘total evidence’ principle of Kluge (1989).



Gpd sequence data (457 bp) were obtained from 110 isolates and used for the initial phylogeny; 103 isolates were from our oak-woodland study site and seven isolates were included from outside of the study site for comparative purposes (Fig. 1). From our oak woodland site, isolates of C. geophilum were obtained from 13, 10, and 16 soil samples from subpopulations one, two and three, respectively. A total of 40, 20, and 43 isolates were obtained from subpopulations one, two, and three, respectively, consisting of 1–6 isolates per soil sample. Three distinct lineages were found within the sampled isolates based on the unrooted NJ tree (Fig. 1). Similar topologies were also observed in MP and ML trees (data not shown). A single soil sample often hosted isolates that were genetically distributed across the phylogeny (e.g. isolates from subpopulation one, sample five or subpopulation three, sample 2). Ten isolates did not possess the CgSSU intron with four of the isolates forming a strongly supported clade with 93% bootstrap (Fig. 1). Forty-six isolates, all possessing an indel (missing) of 42–44 bp in the gpd region, formed a moderately well supported clade with 79% bootstrap support (Fig. 1).

Similar tree topologies were also found for ITS, the CgSSU intron, and mit SSU rDNA regions when the analyses were conducted from the reduced isolate data sets. Fig. 2 shows one of three parsimony trees for gpd with a consistency index (CI) of 0.981, retention index (RI) of 0.995, and length of 54. A total of 498 bp were analyzed with 36 parsimony informative characters, 16 parsimony uninformative characters, and 446 constant characters. MP analysis of ITS produced a single parsimony tree with CI of 0.939, RI of 0.984 and a length of 38. A total of 444 bp were analyzed with 29 parsimony informative characters, one parsimony uninformative character, and 414 constant characters. MP analysis of CgSSU intron produced 10 parsimony trees with a CI of 0.897, RI of 0.978 and a length of 58. A total of 504 bp were analyzed with 47 parsimony informative characters, three parsimony uninformative character, and 454 constant characters. MP analysis of the mit SSU rDNA region produced a single parsimony tree with CI of 1.00, RI of 1.00 and a length of 13. A total of 544 bp were analyzed with 10 parsimony informative characters, two parsimony uninformative characters, and 532 constant characters. When gpd, ITS, and CgSSU intron were combined, three MP trees were found with a CI of 0.895, RI of 0.974 and a length of 152 (Fig. 3). A total of 1446 bp were analyzed with 112 parsimony informative characters, 20 parsimony uninformative character, and 1314 constant characters.

Figure 2.

Maximum parsimony analysis of Cenococcum geophilum isolates from four regions of DNA; glyceraldehyde 3-phosphate dehydrogenase (gpd), ITS-rDNA, a Group I intron located in the 3′ end of the SSU-rDNA and a portion of the mitochondrial SSU-rDNA. Isolates were chosen based on Fig. 1 phylogeny. Isolates from the California oak-woodland are labeled using a three number code: first letter = subpopulation, second = sample, and third = isolate number. Additional non-California isolates are 03-04-II from Maryland oak savannah soil and Btree1, I-2 and I-3 from Oregon (host: Quercus garryana). Bootstrap support of 70% and above are indicated above nodes based on 1000 replicates. Sequences have been deposited in GenBank under accession numbers (AY818546AY818610).

Figure 3.

Maximum parsimony analysis of Cenococcum geophilum isolates from combined data from glyceraldehyde 3-phosphate dehydrogenase (gpd), ITS-rDNA, and a Group I intron located in the 3′ end of the SSU-rDNA. Isolates were chosen based on Fig. 1 phylogeny. Isolates from the California oak-woodland are labeled using a three number code: first letter = subpopulation, second = sample, and third = isolate number. Additional non-California isolates are 03-04-II from Maryland oak savannah soil and Btree1, I-2 and I-3 from Oregon (host: Quercus garryana). Bootstrap support of 70% and above are indicated above nodes based on 1000 replicates.

Three major lineages (I, II, & III) were detected with 98% or greater bootstrap support in the individual analyses of gpd, ITS, and CgSSU intron with additional well-supported minor clades within primarily lineage III. The mit SSU rDNA MP tree also grouped the isolates into the same three major lineages but with less support. In the combined analysis, these three clades had 100% bootstrap support (Fig. 3). In preliminary analyses, we tried to identify a possible outgroup species to use in our study. We found that some of the gpd region (c. 350 bp) was alignable with other Loculoascomycete sequences deposited in GenBank. From this initial alignment, two regions of the gene were conserved, which likely represent conserved coding regions. When an Alternaria species was forced to be the out group taxon, the same three lineages were also found with bootstrap support of 92, 94 and 99% on this reduced data set, including the isolates without the CgSSU intron (data not shown). This analysis also showed that the Cenococcum isolates formed a monophyletic clade, with 100% bootstrap support, apart from five additional Loculoascomycete genera used in the analysis (data not shown). However, this result was from a small data set and additional sequence data from more species and genes would be required to test this more robustly.


Significant phylogenetic divergence among local isolates of C. geophilum was found in this study, suggesting this morphological species is actually a species complex. We detected as much genetic diversity within a single soil sample as was found for isolates collected across the country, even with the limited number of isolates from outside of our oak-woodland used for comparative purposes. Our results may help explain the large amount of physiological, phenotypic, and genetic differences reported among isolates of C. geophilum in many studies (e.g. LoBuglio, 1999) and why previous population genetic studies have concluded that a finer sampling scale was needed to accurately describe the population structure of C. geophilum (Jany et al., 2002; LoBuglio & Taylor, 2002).

Genetic analyses have challenged morphological species concepts within many groups of organisms and have been especially helpful in delineating fungal species that have few morphological characters (Taylor et al., 2000). Taylor et al. (2000) have advocated the use of using the analyses of multiple genes as a criterion to identify phylogenetic species within the fungi, which they term Genealogical Concordance Phylogenetic Species Recognition (GCPSR). They suggest the use of multiple genes to determine the transition from concordance to conflict among taxa, which can be used to determine species boundaries. The conflict is thought to be due to recombination occurring between individuals in a ‘species’ and is usually determined based on incongruence tests, such as the Partition Homogeneity Test (PHT) (= Incongruence Length Difference test, Farris et al., 1995), between genes or DNA regions (e.g. Taylor et al., 2000). Thus, an important assumption of using this approach is that recombination must be able to occur in the species of interest (Taylor et al., 2000), which has been reported for C. geophilum (LoBuglio & Taylor, 2002).

In this study, we found three major lineages within our sampled C. geophilum isolates based on four DNA regions (Fig. 2). A transition from concordance to conflict was not found in lineages I or II but was found within lineage III based on PHT (data not shown). Therefore, within lineage III, this may be further evidence for detecting recombination in C. geophilum. However, we feel that more data and analyses are needed to adequately test this since Dolphin et al. (2000) and Barker & Lutzoni (2002) have shown through computer simulations that interpretation of the PHT test can be questionable. Moreover, the incongruence was only due to the presence of the Oregon isolates I-2 and I-3 in the gpd data set. Therefore, if we accepted that recombination has occurred within lineage III, then it would have likely happened in the history of this lineage and not in the present day local population from our sampled oak-woodland. We are currently investigating the reproductive biology of our sampled population using sequence analyses of additional loci in combination with multilocus genotypic data.

Based on molecular phylogenies, cryptic species have also been suggested for many morphospecies within various fungal genera including Fusarium (Skovgaard et al., 2002), Stachybotrys (Cruse et al., 2002), Tricholoma (Horton, 2002), Coccidioides and some of its close relatives (Koufopanou et al., 2001), and lichenized genera such as Physcia (Myllys et al., 2001) and Letharia (Kroken & Taylor, 2001). In this study, C. geophilum putative cryptic species were found in the same soil sample and from distant geographic regions. Similar results were found by Cruse et al. (2002) for Stachybotrys chartarum, a toxigenic fungus implicated in sick building syndrome. They found cryptic species occupying the same room of a building as well as between states within the United States. With respect to fine scale diversity, our results are similar to that of Moyersoen et al. (2003) and Skovgaard et al. (2002). Moyersoen et al. (2003) found two closely related species of the EM fungus Pisolithus on roots in the same volume of soil. This finding, however, was only possible due to the recent recognition that Pisolithus tinctorius sensu lato consists of at least 11 phylogenetic species (Martin et al., 2002). Skovgaard et al. (2002) found up to three phylogenetic species of Fusarium oxysporum from a single soil sample. They also found no correlation between individuals from each phylogenetic species and the substrate that the isolates were obtained from (soil or diseased pea tissues) or with the pathogenicity of the isolates to pea.

It was important to determine the presence of distinct lineages within C. geophilum because our long-term goal was to study the genetic structure of this ‘species’ using molecular markers. If isolates were arbitrarily pooled as a single species and analyzed by the sampled subpopulation, it could significantly affect the biological interpretation of the data. For example, one may detect gametic disequilibrium among loci if individuals from two separate ‘populations’ or ‘species’ are analyzed as if they are from a single population (population admixture), leading to an erroneous acceptance of a hypothesis of clonal population structure (Milgroom, 1996). LoBuglio & Taylor (2002) pooled isolates of C. geophilum from soil samples collected along transects within conifer forests and detected evidence of a random mating population structure based on analyses of multilocus SNP data. Their results suggest they were likely pooling C. geophilum isolates from the same ‘population’ or ‘species’ since admixture would have likely resulted in the rejection of the null hypothesis of random mating. Likewise, population genetic analyses can also support the existence of reproductively isolated populations or cryptic species such as within the genus Cantharellus as suggested by Dunham et al. (2003) based on microsatellite data.

While we detected distinct phylogenetic lineages among our C. geophilum population, how this genetic diversity relates to phenotypic diversity in the ecology and biology of C. geophilum is unknown. Classic ecological theory predicts that the stable coexistence of identical competitors will not occur (Hardin, 1960) suggesting that cryptic species cannot occupy the same niche and must play different ecological roles. However, alternative theories have also been suggested to account for co-occurrence of species with apparent identical niches (e.g. Zhang et al., 2004). For cryptic species in animals, Ortells et al. (2003) detected the presence of five cryptic species of rotifers in four coastal Mediterranean ponds. They found that a temporal and spatial distribution of the cryptic species was due to ecological specialization, which allowed seasonal succession and partitioning of resources. Pfenninger et al. (2003) demonstrated that environmental data related to climate showed a significant differentiation among divergent lineages (cryptic species) of the freshwater limpet, Ancylus fluviatilis. We suspect that the unique phylogenetic lineages of C. geophilum are in fact occupying unique niches within our ecosystem. We are currently testing the hypothesis that the different putative cryptic species of C. geophilum will correlate with unique physiological profiles, which should give us an indirect indication that they may be occupying unique niches. It is also possible that not all fungi with the C. geophilum morphology are in fact mycorrhizal; this would be analogous to pathogenic and nonpathogenic isolates of F. oxysporum that have been found occurring in the same habitats (Skovgaard et al., 2002). However, we have identified all three lineages of C. geophilum directly from colonized oak roots from our study site, demonstrating that all three lineages are mycorrhizal (G. W. Douhan, unpublished).

C. geophilum sensu lato is clearly geographically widespread and ecologically successful, which is amazing given its apparent asexual nature and inability to produce any type of spore for dispersal. Recognizing that C. geophilum sensu lato is a species complex may help to explain the apparent success of this single ‘species.’ However, further clarification of the broad scale phylogeny and fine scale population genetic structure of C. geophilum sensu lato throughout its known range is needed in order to understand the phylogeography of this ubiquitous morphospecies complex. Detailed biological studies may then reveal associated phenotypic differences (morphological, physiological) between putative cryptic species within C. geophilum that may lead to a better understanding of the ecology of the mycorrhizal symbiosis. The spatial scale at which we have found divergent lineages suggests that further work should yield results that may have broader implications in the study of fungal ecology.


We thank Jim Trappe & Danniel Panaccione for C. geophilum cultures, Darlene Southworth for sclerotia samples, and Marc-André Selosse, Kathy LoBuglio and anonymous reviewers for helpful comments and suggestions made during the editorial process of this manuscript. We also thank Mark White, Karyn Huryn, and LeAnn Douhan for technical help. This study was supported by the National Science Foundation Biocomplexity program (DEB 9981711, DEB 9981548).