Author for correspondence:Thomas D. Bruns Tel: +1 510 6427987 Fax: +1 510 6424995 E-mail: firstname.lastname@example.org
• Our primary goal was to determine whether detection of Rhizopogon internal transcribed spacer (ITS) groups is affected by the pine species used in seedling bioassays. Our secondary goal was to investigate composition of Rhizopogon spore banks in the Eastern Sierra Nevada of California, a previously unsampled region.
• We used seedlings of Pinus contorta, Pinus jeffreyi, Pinus lambertiana, and Pinus muricata as bioassay plants and identified the Rhizopogon retrieved by internal transcribed spacer (ITS) sequence analysis.
• We found that each of the pine species retrieved all of the abundant Rhizopogon ITS groups, but there were significant differences among pines in the richness of Rhizopogon ITS groups recovered. Pinus muricata recovered all ITS groups found in this study and was significantly better than P. lambertiana. Rhizopogon communities from the five sampled sites contained six to eight ITS groups per site, with two unique sequence groups and a higher abundance of the Rhizopogon ellenae and Rhizopogon arctostaphyli groups than at previously sampled sites.
• These results show high cross-receptivity between Rhizopogon and pine species, and regional patterns in spore bank composition.
Ectomycorrhizal fungi are necessary for the growth and survival of pines and many temperate forest species (Smith & Read, 1997). These fungi are essential for acquisition of mineral nutrients and can protect roots from adverse conditions such as drought and high metal concentrations (Brownlee et al., 1983; Burgess et al., 1993). Rhizopogon is a genus of ectomycorrhizal fungi with over 250 described species and varieties, and most of those species are specifically associated with members of the Pinaceae (Molina et al., 1999). Recent work by Kjøller and Bruns (2003) has shown that Rhizopogon species form abundant spore banks in pine-dominated forests in California, and these spore banks can be efficiently sampled by bioassaying diluted soil with pine seedlings. Following severe disturbance, such as stand-replacing fire, these spore banks provide most of the mycorrhizal inoculum for newly establishing pine seedlings (Horton et al., 1998; Baar et al., 1999).
Our primary goal was to test the hypothesis that different pine species yield different species of Rhizopogon under identical bioassay conditions. This question has practical implications, because some pine hosts are much faster and easier to grow than others, and if one were to design an efficient survey of Rhizopogon based on pine bioassays, it would be important to know which pines provide the best assay plants. The question is also important for biological reasons, because if all pines can interact with the same Rhizopogon spore banks it could facilitate migrations of both pine and Rhizopogon species in diverse floristic settings such as California where different pine species typically dominate at different elevational and latitudinal zones (Barbour & Major, 1990).
It is well established that some Rhizopogon species are specific to either Pinus or Pseudotsuga, and that host specificity within Rhizopogon correlates with the phylogeny of the Pinaceae (Massicotte et al., 1994; Molina et al., 1999; Grubisha et al., 2002). However, Rhizopogon specificity within the genus Pinus has not been well investigated. Suillus, the sister genus to Rhizopogon, has been observed to discriminate between hard pines (subgenus diploxylon) and soft pine (subgenus haploxylon) (Smith & Thiers 1964), and there is some evidence that Rhizopogon may also discriminate. For example, two species are known only from single hosts in subgenus haploxylon. Rhizopogon pinyonensis is so far known only from pinyon pine (Harrison & Smith, 1968) and an apparently undescribed species of Rhizopogon was retrieved from bioassays with Pinus longaeva (Bidartondo et al., 2001). However, most collections of pine-associated Rhizopogon appear to be associated with hard-pines (J. M. Trappe, pers. comm.; T. D. Bruns pers. obs.).
Based on these observations we hypothesized that some Rhizopogon species would discriminate between hard and soft pines, but that the majority of species would not discriminate among hosts within a subgenus. To test this hypothesis, we collected soils from five sites and bioassayed it with Pinus lambertiana, Pinus jeffreyi, Pinus contorta and Pinus muricata and then determined the taxa of Rhizopogon retrieved from each host. The first three hosts are all native to the Sierra Nevada, while the last is restricted the coastal California. Pinus lambertiana is a soft pine, while the three others are hard pines. Pinus muricata can be germinated within a week but the three montane species require longer vernalization treatments for efficient germinate. For these reasons, P. muricata is faster and more convenient to use, and it was the only host used for the previous bioassay survey of coastal and Western Sierra Nevada sites (Kjøller & Bruns, 2003).
Our second goal was to sample composition of Rhizopogon spore banks from sites in the Eastern Sierra Nevada as part of a long-term study of Rhizopogon distributions throughout the California floristic province. Previous bioassay studies of pine forests were limited to four sites on the coast dominated by P. muricata, one site at moderate elevation in the Western Sierra Nevada dominated by Pinus ponderosa, and one site at high elevation in the California White Mountains dominated by P. longaeva (Bidartondo et al., 2001; Kjøller & Bruns, 2003). From these studies it was determined that the composition of Rhizopogon spore banks was similar in the Northern coastal sites and the Western Sierra site, but was very different at the Southern most coastal site and in the White Mountains. The Eastern Sierra sites sampled in the current study contain a different set of pine species, a different climate, and different soils than the other sites sampled, and for these reasons we expected to see differences in the composition of Rhizopogon spore banks.
Materials and Methods
Pinus l. Douglas, P. jeffreyi Grev. & Balf. and P. contorta Loudon seeds were provided by the Placerville Nursery (USDA Forest Service). Pinus muricata D. Don cones were collected from Point Reyes National Seashore and the seeds were harvested directly from the serotinus cones. All seeds were surface-sterilized in 30% hydrogen peroxide with a drop of Tween-20, for 20 min, and then rinsed with water. P. lambertiana, P. jeffreyi and P. contorta seeds were vernalized, as suggested by the Forest Service Nursery in the following way. Seeds were placed in individual plastic bags with small holes punctured in the bottom and placed under running water for 2 d followed by a 30-d drip-dry period in the cold room (4°C). Seeds were covered by moist paper towels in the plastic bags to retain moisture and were checked every other day to make sure the moist environment was maintained. After 30 d the seeds were rinsed, again, drip-dried, and placed in a new plastic bag and in the cold room for the following lengths of time: P. jeffreyi, 33 d; P. lambertiana, 61 d; and P. contorta, 15 d. Vernalization was timed so that seeds for all species would be ready to plant at approximately the same time. This process was started on 15 July 2003 and ended on 14 October. The P. muricata seeds did not require any soaking or drying time and were ready to plant immediately after they were surface sterilized and rinsed.
Soil samples were obtained from Eastern Sierra pine sites near Mammoth Lakes area between 18 and 20 June, 2003 from the following five sites: Obsidian Dome (37°45.403′, 119°1.456′, altitude 2588 m); Forest near Obsidian Dome (37°45.433′, 119°1.652′, altitude 2573 m); Inyo Crater Area (37°41.277′, 119°0.370′, altitude 2484.6 m); Fern Lake Trail Head (37°45.754′, 119°6.812′, altitude 2242 m); and Ribbon Falls Area (37°36.194′, 119°4.904′, altitude 2287 m). The first three sites had soil derived from volcanic activity within the last 600–900 yr. Ribbon falls was a young postfire forest derived from a burn less than 10 yr old. Pinus jeffreyi was the dominant pine at all sites except the Fern Lake trailhead, where P. contorta was more common. Pinus contorta was also found at Ribbon falls, and at the forest near Obsidian Dome. The latter site also contains Pinus monticola, a soft pine, and Abies concolor. Each soil sample consisted of 10 small scoops of soil, approx. 10 cm into the mineral layer. Each scoop was collected 2–3 m from the previous scoop along a haphazardly placed line. To avoid cross-contamination the trowel was wiped clean with an ethanol-soaked paper towel between sites.
Each of these five sample soils was sifted in a no. 10 soil screen to remove large rocks, roots, and other debris from the samples. This process was carried out in a transfer hood to prevent contamination between soils. All sifting equipment was soaked in a bleach solution between siftings for 10 min, rinsed in water for 5 min, and then dried in an oven. Each sifted soil was transferred to plastic bags and stored in the cold room until needed. Extra soil is archived at the Department of Plant and Microbial Biology, University of California at Berkeley.
Soil collected from Tomales Point, in a region known to be devoid of Rhizopogon (T. D. Bruns, unpublished) was autoclaved for 60 min with a 10-min drying cycle. A total of 300 ml of this sterile soil and 300 ml of sterile coarse sand was mixed with 4.5 ml of the inoculum soil in a sterile quart mason jar (approx. 946 ml), producing an approximate volume of 600 ml per jar. Separate sterile jars were used for each test soil. This dilution step was used because previous studies have shown it to be an efficient way to retrieve Rhizopogon from soil (Taylor & Bruns, 1999; Kjøller & Bruns, 2003). Each jar provided enough inoculated soil to fill 12 bioassay tubes (RLC-4 Super ‘Stubby’ Cell Container; Stuewe & Sons Inc., Corvallis OR, USA) to a volume of 50 ml. Polyester fiber was stuffed to an approximate height of 2–3 cm to prevent any soil added from falling through the drainage slots in the bottoms of the tubes, as well as to allow for sufficient drainage while maintaining a moist environment. For each site 40 tubes were set up for each of the four test pine species (n = 160 tubes per site); this number was selected because in an earlier study 20–40 Rhizopogon samples were found to saturate the accumulation curve for a site (Kjøller & Bruns, 2003). In addition 160 uninoculated controls (40 per host) were set up with only sterile Tomales Bay soil. Tubes were watered to moisten the soil and two or three pine seeds were distributed into each tube using sterilized forceps. These were later weeded to achieve one seedling per tube. Approximately 1 cm of sterilized sand was added on top of the seeds. A total of 960 tubes were planted in this way: 5 soil types × 4 hosts × 40 replicates + 160 uninoculated controls.
Isolation of Rhizopogon
Seedlings were harvested 5–6 months after planting. Suilloid ectomycorrhizal root tips were identified under a dissecting scope by their clustered morphology, cottony white surfaces, and thick white-branched rhizomorphs (Agerer 1987–2002; Kjøller & Bruns, 2003). Typically one or two mycorrhizal tips were collected from each seedling. This small sampling within seedlings was chosen because we have found that at high levels of sample soil dilution seedlings are typically colonized by one species of Rhizopogon. Individual tips were placed in 2 ml Eppendorf screw-top tubes containing three glass beads, snap-frozen in liquid nitrogen, and stored in a −80°C freezer until they were freeze-dried. Individual freeze-dried root tips were crushed in a bead beater and DNA extraction was performed using the Dneasy Tissue Kit (Qiagen Inc., Valencia, CA, USA). Polymerase chain reaction (PCR) amplification was carried out using the internal transcribed spacer (ITS) of the fungal ribosomal DNA with the primers ITS1-F and ITS4, as described by Gardes & Bruns (1993). We initially screened PCR amplicons by restriction fragment length polymorphism (RFLP) analysis to reduce sequencing. Separate aliquots of PCR products were digested with the enzymes Hinf-I and Hha-I (purchased from New England Biolabs Inc., Beverly, MA, USA), and restrictions digest patterns were visually compared on agarose gels. Multiple samples of all unique RFLP patterns from each host–site combination were sequenced. Sequencing reactions were performed following instructions from the BIGDYE v3.1 chemistry (Applied BioSystems, Foster City, CA, USA) and reactions were run on an ABI 3100 automated sequencer (Applied BioSystems, Foster City, CA, USA). Data obtained were processed using DNA sequence analysis (version 2.01) and sequence navigator software (Applied BioSystems, Foster City, CA, USA).
Sequences analysis and phylogenetic placement
We used phylogenetic analysis to place unknown sequences into terminal groups that contained low levels of nucleotide divergence and no discernible phylogenetic structure. We did this in five steps. (1) We used blastn to screen out sequences out were not Rhizopogon. (2) We assembled a baseline ITS dataset using previously deposited sequences of Rhizopogon derived from identified basidiocarps and from unidentified root samples. These were retrieved from GenBank via Entrez by targeting sequences deposited from several recent papers (Bidartondo et al., 2001; Bidartondo & Bruns, 2002; Grubisha et al., 2002; Kjøller & Bruns, 2003; Kretzer et al., 2003) and by using the key word search: ‘Rhizopogon’ and ‘internal transcribed spacer’. Retrieved sequences containing only one of the two spacer regions were discarded and identical sequences were dropped. The final baseline data set consisted of 139 Rhizopogon sequences derived from GenBank and 113 derived from this study. Accession numbers for the sequences used and produced in this study are given in the supplementary table (Table S1). All of the sequences were then aligned using POA (Lee et al., 2002). We trimmed this alignment at the 5′ and 3′ ends to the point where most of the sequences contained data, but the alignment was otherwise unaltered. (3) We used neighbor-joining (NJ) analysis as implemented in paup (Swofford, 2002) to identify groups of highly similar sequences. This ‘broad’ analysis was used to place the unknown sequences into smaller groups of more similar sequences. (4) We aligned sequences from each group identified by the broad analysis and their closest relatives with POA and analysed these groups separately with parsimony in ‘group-specific’ analyses. We used heuristic searches using 10 random addition sequences, and set the maxtrees options to 10 000. We then examined the strict consensus tree to look for common internal structures, and several individual trees to visually exam branch lengths. (5) We used Jacknife analysis as implemented in paup with 33.3% deletion, and the fast search option, to examine support for internal branches.
Differences in the number of seedlings colonized and ITS group richness among hosts and sites were compared using a series of one-way anovas. The ITS group richness was analysed using both the raw data as well as richness estimates based on rarefaction (Krebs, 1999) to correct for differences in the number of seedlings colonized across hosts. For host comparisons, ITS group richness estimates were rarefied to five seedlings per host per site, which was the minimum number of seedlings colonized on any host (at one site, P. lambertiana had only three out of 40 seedlings colonized, but this sample was eliminated from statistical analyses of host effects). When comparing ITS group richness among sites, the data were first rarefied within each host (n = 3–8 depending on host) in order to minimize the differences based on hosts at each site. Post-hoc comparisons of hosts and sites were conducted using Tukey HSD tests. Before each anova, variances were determined to be homogeneous using Levene's test (P > 0.05). The ITS group richness at each site in this study and that in Kjøller & Bruns (2003) were compared using a two-tailed t-test following rarefactions to standardize sample sizes (n = 8) and limiting comparison to assays with P. muricata. The variances were determined to be homogeneous using an F-test. The occurrence of the Rhizopogon occidentalis group across hosts was tested using a χ2 test of independence. All analyses were conducted in jmp 5.0 (SAS Institute Inc, Cary, NC, USA) and considered significant if P = 0.05. Bootstrap estimates of Rhizopogon ITS group richness were calculated for each site using the pooled data from all four hosts and estimates software (Colwell, 2005).
A total of 295 of the 800 test seedlings were visually selected for analysis based on suilloid-like morphologies. Rhizopogon was confirmed in 285 of these, the other 10 yielded sequences of Suillus (6), Phialophora (2), Hebeloma (1), and an unidentified basidiomycete (1). In addition five seedlings contained two different Rhizopogon ITS groups for a total of 290 Rhizopogon positive samples. No Rhizopogon was found on any of the 160 control seedlings. We recovered 11 Rhizopogon ITS groups from the positive bioassays. These groups were based on 113 sequences (Table 1), which represented multiple samples of all RFLP types from each site and usually from multiple hosts. Nine of the groups defined are the same as those recognized in previous study of Rhizopogon spore banks in other parts of California, or from roots of the Monotropoideae (Bidartondo & Bruns, 2002; Kjøller & Bruns, 2003), but two are unique to this study. One of them, gr. 10, is placed in section Amylopogon as a unique clade with high confidence (several branch > 95% jackknife), and is placed as the sister group to group 8 (arctostaphyli) with low confidence (not shown). Group 16, the other unique group, is placed in section Rhizopogon in the neighborhood of a sequence identified as Rhizopogon subsalmonius, but it is not particularly close to that sequence (not shown). Most of the sequence groups defined in Table 1 contain few sequence differences, no discernible within-group phylogenetic structure and are well separated from other groups in the phylogenetic analyses conducted. Groups 4a and 4b are exceptions to these rules in that they contain more within-group divergence, and differences between the two groups are not well supported. However for the purposes of searching for possible host discrimination among different Rhizopogon taxa, we think it is better to divide groups as finely as possible.
Table 1. Summary of internal transcribed spacer (ITS) sequence groups
Total number of sequences analysed (number derived from this study).
Mean (and maximum) within group uncorrected substitutions calculated as the number of bases that differ divided by the number of aligned bases.
na, not applicable.
gr 4a roseolus
gr 4b vulgaris
gr 5 ellenae
gr 8 arctostaphyli
gr 9 salebrosus
gr 10 sect. Amylopogon
gr 11 occidentalis
gr 13 fuscorubens
gr 14 evadens
gr 15 olivaceotintus
gr 16 sect. Rhizopogon
Differences in the distribution of the 11 ITS groups were evident across both hosts and sites (Fig. 1, Table 2). The number of seedlings colonized was not significantly different among hosts (F = 0.577, P = 0.638), however, ITS group richness was significantly different (F = 3.333, P = 0.046). Pinus muricata seedlings recovered greater richness of Rhizopogon ITS groups than P. lambertiana, with P. jeffreyi and P. contorta being intermediate (Tukey test, P < 0.05) (Table 3). The ITS group richness pattern was even stronger when controlling for differences in sample size across hosts using rarefaction (F = 6.137, P = 0.006). Both P. muricata and P. contorta seedlings had significantly higher ITS group richness than P. lambertiana, while ITS group richness of P. jeffreyi was not significantly different from any other host (Tukey test, P < 0.05). The number of seedlings colonized also varied significantly among sites (F = 10.741, P < 0.001). Over half the seedlings from both Inyo Crater and Fern Lake trailhead were colonized, while less than a quarter of the seedlings were colonized from the forest near Obsidian Dome, Obsidian Dome, and Ribbon Falls (Table 4). The observed ITS group richness appeared to vary between sites (Table 4), but it was not significant when sample sizes were controlled within hosts (F = 0.11, P = 0.973). Accumulation curves for Rhizopogon ITS groups appeared to plateau for three of the sites when hosts were pooled (data not shown). Bootstrap estimates of Rhizopogon richness were only slightly greater than the observed values for all five sites, and they indicated that at most one ITS group may have been missed at Obsidian Dome and Ribbon Falls sites (Table 4).
Table 2. Rhizopogon internal transcribed spacer (ITS) sequence groups found across sites and host
Proportion of seedling colonized by Rhizopogon averaged across the four hosts; SE in parenthesis.
Number of Rhizopogon ITS groups recovered by all four hosts.
Bootstrap estimate of richness; SD in parenthesis.
Obsidian dome forest
Fern Lake Trailhead
When we examined the recovery of individual Rhizopogon ITS groups, we found no obvious differences between hosts. Six of the 11 Rhizopogon sequence groups were found on all four pines within at least one site (Table 2), and these six included all of the abundant ITS groups. The five groups that were not found on all potential hosts were the rarest: (1) Rhizopogon olivaceotinctus group was recovered from eight seedlings across two sites and was not found on P. contorta; (2) Rhizopogon fuscorubens group was recovered on three seedlings across two sites was not found on P. jeffreyi and P. lambertiana; (3) Amylopogon group 10 was recovered from four seedlings across two sites but was not found on P. contorta or P. lambertiana; (4) group 16 was recovered from five seedlings at a single site, but was not found from P. jeffreyi; and (5) R. occidentalis group was recovered on 13 seedlings across four sites, but was not found on P. lambertiana. Of these examples only the R. occidentalis group appeared to be abundant enough that its absence on P. lambertiana may not have been caused by chance. However, the observed pattern was not significantly different from a random pattern ( = 11.08, P = 0.271). In addition, we know from inoculation studies that R. occidentalis has the ability to colonize P. monticola (Molina & Trappe, 1982), which, like P. lambertiana, is in the soft pine group. Interestingly, all Rhizopogon ITS groups were recovered by P. muricata, the only host species tested that is not native to the Sierra.
While the Rhizopogon spore banks across the five sites differed in terms of the relative abundance of individual ITS groups (Fig. 1), their higher-level composition and general structure were similar (Fig. 2). Three ITS groups were found in all five locations: Rhizopogon arctostaphyli, Rhizopogon ellenae and Rhizopogon roseolus, and these were often among the most abundant groups. Only group 16 was unique to one site, even though several other groups were consistently low in abundance. At all sites > 80% of the Rhizopogon-positive seedlings were colonized by a combination of species in the subgenus Amylopogon (groups 5, 8–10) and the roseolus–vulgaris clade (gr. 4a and 4b), while species in the remaining five groups were low in abundance (Table 2, Fig. 1). Ignoring differences in taxon identity, the three most abundant ITS groups at a site accounted for 79 ± 2% of the colonized seedlings, and the structure of the communities fit a log normal distribution (Fig. 2).
All four pine species tested appeared to have the ability to associate with all of the Rhizopogon ITS groups encountered. We only qualify this statement because particular hosts did not recover several rare types. These missing Rhizopogon–pine combinations are probably explained by chance, but because these ITS groups were rare in our sample, we did not have the ability to detect species-specific effects involving them. There could be different biases among hosts for particular taxa of Rhizopogon, but if such biases exist they are not obvious (Table 2). The fact that no Rhizopogon ITS groups appeared to be specific to the hard or soft pine subgenera is very different from the pattern reported in Suillus, where many species that are distinct at the ITS level are associated with either hard or soft pines but not both (Kretzer et al., 1996). Our results with Rhizopogon mean that there is broad cross-receptivity among pines; this could be important as pine-dominated vegetation zones move with changing climate or changing fire regimes.
Pinus muricata, the only species not native to the Sierra, actually performed the best, exhibiting the highest level of Rhizopogon-positive bioassays, recovering the greatest average richness of Rhizopogon across all sites, and missing none of the ITS groups detected in our study. It also recovered significantly higher levels of Rhizopogon richness than P. lambertiana. This result has practical importance for seedling bioassay studies, because the seeds of P. muricata are more easily germinated than many of the montane species (see the Materials and Methods section).
The host specificity of Rhizopogon species has been extensively studied among genera of the Pinaceae and arbutoid hosts in the Ericaceae (Massicotte et al., 1994, 1999; Molina & Trappe, 1994). From these studies it is known that Rhizopogon specificity varies with context of the assay: syntheses based on mycelial cultures, or dual-host pots exhibit broader host ranges than those based on spore inoculations with single hosts. In the case of mycelial cultures, this may be an artifact of sugar level in the media (Massicotte et al., 1994). However, in the dual-host setting, the lower apparent specificity likely reflects the ecological potential in which a species that has already colonized its primary host by spore may be able to expand via mycelium onto a secondary host (Massicotte et al., 1994). This is relevant to the current study in that single host–spore inoculation assays are the most restrictive setting, yet our single host assays showed no specificity among pine species.
Composition compared among sites in this and other studies
The species richness observed at these Eastern Sierra sites, was slightly higher than that reported on Western Sierra and Coastal sites by Kjøller and Bruns (2003). Excluding their southern coastal site, which was a clear outlier in the earlier study, the three central and northern coastal sites and the Western Sierra site had four to six ITS groups (mean = 5), while the Eastern Sierra sites, varied from five to eight groups (mean = 6.8). The confounding factors of slightly different methods, spatial samples and sample sizes make quantitative comparisons between the two studies difficult. However, ignoring the difference of slightly greater within-site spatial scale of the earlier study (which should bias it toward greater richness), and the culture bias of that study (which may have reduced richness) there was no significant difference in group richness (T = −1.75 df = 7, P = 0.123). Despite the fact that differences were not significant, it is clear that the Eastern Sierra sites are at least as species rich as the coast and Western Sierra sites.
The high richness of Rhizopogon in the Eastern Sierra sites is interesting because some of the sites are amazingly harsh. Obsidian Dome, for example, is an 600-yr-old volcanic dome of unvegetated rhyolite and obsidian with scattered islands of pine that are separated from the adjacent forest by several hundred meters of bare rock. Our sample from this site included four pine islands, three of which consisted of only one or two trees. Obtaining sufficient soil for these samples was difficult; yet, this site had the highest richness (Fig. 1, Table 4). The adjacent Obsidian Dome forest and the Inyo Crater forest were less harsh, and contain fairly continuous, open forest, but the soil of both was derived from volcanic activity of about the same age as Obsidian Dome. Izzo et al. (2005) has shown that even though some of the Rhizopogon species are not resident colonizers or sporocarp producers in mature forests, they are present in spore bank samples. In this way they are similar to some post-fire plant species that reside in a seed bank between disturbance events. Thus, the richness of Rhizopogon on these harsh sites may simply be a function of their early successional status, or perhaps the drier climate of the Eastern Sierra, which may favor dispersal by mammals over wind dispersal, or greater persistence in the spore bank.
The composition of these Eastern Sierra communities contained some major differences from the coastal sites. The high frequency of the R. arctostaphyli and R. ellenae groups across the Eastern Sierra sites was one striking difference, as the former group was absent from the coast and the latter was restricted to the Northern most coastal site. Similarly, the R. roseolus group, which was found at all five sites in this study, was found at only one of the coastal sites. There were also two novel groups found: group 10, which is related to the R. arctostaphyli clade, but distinct from it, and group 16, which is unique relative to all sampled taxa within the genus. Neither of these new groups was particularly abundant, unlike the novel Amylopogon clades found at the southern coastal sites (Kjøller & Bruns, 2003).
The effects of soil dilution on perception of the Rhizopogon spore banks
The purpose of soil dilution before bioassay is to limit competitive effects among fungal species so that a few good competitors do no dominate the assay and mask the presence of lesser competitors. To achieve this goal, the soil must be diluted enough so that most seedlings are colonized by a single species. The low colonization levels observed (Table 4) are thus in line with this goal, but at the cost of sample size, since the majority of bioassay seedlings yielded no data. There is also the risk of diluting less abundant species to extinction. The differences in percentages of seedlings colonized by Rhizopogon observed between sites (Table 4) probably reflect differences in spore concentration. The alternative hypothesis that differences in soil properties affected colonization is less plausible as test soils were diluted 134 : 1 into a common Rhizopogon-free soil for all assays. Had we known in advance that some sites had lower spore densities it would have made sense to dilute them less and thereby increase sample sizes.
Even if different dilutions recover the same species, they are likely to provide rather different views of abundance. The most dilute samples will reveal primarily the most abundant species, but the least dilute samples will reveal some mix of the most abundant and the most competitive of the low abundance species. Evidence of this effect was present in the Taylor & Bruns (1999) study that showed a significant increase in number of seedlings by Rhizopogon salebrosus (sect. Amylopogon), a species that we now know can be outcompeted in the initial colonization of seedlings by R. occidentalis (Kennedy & Bruns 2005), and possibly other taxa. Comparison across studies also suggests that species in section Amylopogon increase with dilution (Table 5). In the spore banks at Teakettle forest in the Western Sierra, Izzo et al. (2006) used undiluted soils and recovered a much lower frequency of groups from section Amylopogon compared with the current study or the Kjøller & Bruns (2003) study. The Izzo et al. (2006) study also found a much higher frequency of the R. olivaceotinctus group, which is rare in the other studies cited (Table 5). Differences in the Rhizopogon communities present among the sites could obviously be a factor, but one of the sites in the Kjøller & Bruns (2003) study was only a few miles away from Teakettle Forest and yet the percentage of seedling colonized by Amylopogon taxa and olivaceotinctus gr. were 71% and 2%, respectively, vs 23% and 20% at Teakettle. These comparisons point out that abundance estimates are relative measures that are comparable only among studies that use similar dilutions, and that absolute quantification will be difficult to achieve owing to the confounding factor of competition.
Table 5. Abundance differences in two Rhizopogon internal transcribed spacer (ITS) sequence groups across studies using different soil dilutions
We thank Tim Szaro for his assistance with teaching the senior author the molecular techniques necessary for this work. We thank Inyo National Forest and Devil's Pile Post National Monument for cooperation with the fieldwork. Funding was provided by NSF grant DEB-0236096 to T.D. Bruns, and an REU supplement to T.A. Rusca.