EMF of nursery and glasshouse seedlings
As expected, the taxa colonizing seedlings in the nursery and glasshouse treatments were EMF which have been shown to readily colonize seedlings from spores or soil propagules (Bledsoe & Tennyson, 1982; Castellano et al., 1985; Berch & Roth, 1993; Massicotte et al., 1994; Parlade et al., 1996; Horton et al., 1998; Baar et al., 1999; Taylor & Bruns, 1999). In mature bishop pine (Pinus muricata) forests in California, Taylor & Bruns (1999) found colonization of bioassay seedlings to occur primarily by Rhizopogon spp., T. sublilacina, Tuber spp. and Wilcoxina spp., similar to the genera we observed in our bioassay.
The nursery taxon R. rudus persisted on glasshouse seedlings potted in field soils. However, W. rehmii matching those from the nursery occurred at a much higher abundance on glasshouse seedlings than on field seedlings. We cannot rule out the possibility that at least some of the colonization by Wilcoxina occurred because of spore deposition in the glasshouse, since Wilcoxina is a common glasshouse contaminant. Wilcoxina may be a poor competitor, flourishing only in the absence of the diverse fungal inoculum present in the field. Also, the relatively poor mycorrhizal development of the potted seedlings may have allowed the ectendomycorrhizal Wilcoxina to be detected more frequently by molecular methods. On field seedlings, Wilcoxina could have been present as a secondary symbiont of tips colonized by ectomycorrhizal taxa forming a well-developed mantle, causing the molecular signal to be obscured during PCR amplification because of the more prevalent DNA from the ectomycorrhizal symbiont.
Baar et al. (1999) found that both Rhizopogon spp. and Wilcoxina spp. established readily from resistant soil propagules (i.e. spores) on bishop pine bioassay seedlings. However, Rhizopogon spp. were more common on bioassay seedlings than field seedlings while Wilcoxina spp. were less common, in contrast to what we observed. It should be noted that our glasshouse treatment differed from a bioassay in that seedlings were previously colonized by nursery EMF taxa. The higher relative abundance of Rhizopogon spp. on field seedlings compared with seedlings potted in field soils in our study was intriguing considering the ease with which Rhizopogon spp. colonize roots from spores (Castellano et al., 1985; Parlade et al., 1996; Baar et al., 1999; Kjoller & Bruns, 2003). Rhizopogon spp. may have been favored under dry field conditions (Parke et al., 1983). Glasshouse seedlings differed from field seedlings in that they were irrigated and therefore probably experienced little water stress.
Thelephora terrestris was absent from nursery seedlings, but was common on glasshouse seedlings potted in field soils. Because T. terrestris is a common glasshouse contaminant, colonization of glasshouse seedlings could have occurred either from spores present in field soils or from local airborne spores. Roth & Berch (1992) found that Douglas fir nursery seedlings were dominated by T. terrestris, however, Castellano & Molina (1989) found that presence of T. terrestris varied from nursery to nursery. In our study, there was a trend of higher abundance for T. terrestris on seedlings > 16 m from trees than on seedlings < 6 m from trees. Other studies have shown that T. terrestris decreased in abundance on seedlings near forest edges compared with seedlings further into clearcuts (Durall et al., 1998; Kranabetter & Wylie, 1998; Kranabetter, 1999; Jones et al., 2002). Simard et al. (1997) observed a sixfold increase in Thelephora on trenched Douglas-fir seedlings, implying that forest edge effects on T. terrestris abundance may be caused by inability to outcompete EMF taxa linked to mature trees. Kranabetter & Friesen (2002) found that T. terrestris was able to replace EMF taxa from the forest when western hemlock seedlings were transplanted into forest openings.
EMF of excavated mature tree roots
The genera Tylospora (particularly Tylospora I), Tomentella (particularly Tomentella sublilacina and Tomentella stuposa), Russula (particularly R. nigricans), and Boletus zelleri occurred more frequently on seedlings < 6 m from mature trees than on seedlings > 16 m from trees. These EMF taxa were also common on mature tree roots, with the exception of T. sublilacina. From this pattern it appears that the predominance of T. sublilacina on seedlings < 6 m from trees was probably caused by differences in the soil environment, while the other four taxa may have colonized seedlings directly from mature tree roots. Boletus zelleri was the only EMF taxon forming long rhizomorphs, which would make it a particularly compelling candidate for dependence on hyphal linkages to mature trees.
Boletus zelleri was more frequent and more abundant on seedlings planted in microsites with buried wood compared with sites lacking buried wood, while other EMF taxa did not appear to be influenced by the amount of buried wood (Cline, 2004). Other studies have observed differences in EMF species composition when wood and soil substrates are compared (Kropp, 1982; Goodman & Trofymow, 1998; Smith et al., 2000; Tedersoo et al., 2003). Buried soil wood may be particularly important during periods of drought because of its ability to retain moisture.
EMF of field seedlings
Thirty-four EMF taxa were observed on Douglas-fir seedlings planted > 16 m from mature trees, while 43 taxa occurred on seedlings planted within 6 m of mature trees, using sequence and RFLP analysis. We know of no comparable studies for Douglas-fir. However, these data fall within the range of morphotypes reported for seedlings near forest edges or on isolated trees in studies of other tree species. Naturally regenerated paper birch seedlings near paper birch trees had 47 morphotypes (Kranabetter, 1999), while lodgepole pine, white spruce and subalpine fir seedlings planted near forest edges averaged 52 morphotypes per tree species (Kranabetter et al., 1999). Kranabetter & Wylie (1998) found 44 morphotypes on naturally regenerated western hemlock seedlings in 4-yr-old forest openings near forest edges. Only 25 morphotypes were observed on forest seedlings transplanted into clearcuts, while forest seedlings transplanted into the forest had 38 morphotypes (Kranabetter & Friesen, 2002). Simard et al. (1997) found that trenching reduced the number of morphotypes observed from 17 to 9 for Douglas fir seedlings growing in mature mixed forests of Douglas fir and paper birch. By comparison, Roth & Berch (1992) observed 33 morphotypes on Douglas-fir seedlings 1 yr after outplanting into clearcuts on Vancouver Island.
We found that the EMF community was more species rich for seedlings planted < 6 m compared with > 16 m from mature trees. Diversity indices and rank abundance curves reflected the greater evenness of the EMF community of seedlings < 6 m from trees, although the rank abundance curves for each seedling type were very similar by spring of 2000, 2 yr after planting. Studies of seedlings (Hagerman et al., 1999b) and ectomycorrhizal sporocarp production (Sparks, 2003) near forest edges provide evidence that 16 m is sufficient to isolate seedlings from mature trees. Nevertheless, we are not able to rule out the possibility that roots or, more likely, mycelial strands connected to mature tree roots could have grown out to reach the > 16 m seedlings by the end of the study period.
The relative magnitude of the effect of proximity to isolated mature Douglas fir trees on EMF diversity of Douglas fir seedlings was comparable to that observed in several studies of seedlings growing near mature trees (Kranabetter & Wylie, 1998; Hagerman et al., 1999b; Kranabetter, 1999). The effect was less extreme than the nearly twofold increase in EMF species richness observed for red oak seedlings growing near oak trees (Dickie et al., 2002) and for untrenched Douglas-fir seedlings compared with trenched seedlings growing in mixed Douglas-fir forests (Simard et al., 1997). By contrast, Durall et al. (1998) found only slight decreases in EMF species richness of western hemlock and lodgepole pine seedlings with distance from the forest edge, while Jones et al. (2002) found no differences in species richness with distance for previously colonized Engelmann spruce seedlings planted in mineral soil exposed by mounding. Jones et al. (2003) proposed that forest edge effects on EMF species richness were minimal for previously colonized (i.e. standard nursery) seedlings, but we detected effects of proximity to trees despite the fact that our seedlings were previously colonized. Our seedlings retained their nursery EMF taxa after planting, but significant differences in EMF species richness and diversity developed among seedlings < 6 m compared with > 16 m from residual trees.
Other mycorrhizal trees and shrubs at the site could have provided an additional reservoir of EMF inoculum, although we did not study this. Western hemlock advance regeneration was common at our sites (Cline, 2004). There appears to be some overlap between western hemlock and Douglas fir EMF communities (Roth & Berch, 1992; Smith et al., 1995), a topic that deserves further study. In other western USA forests, Horton & Bruns (1998) showed that a substantial proportion of EMF taxa were shared on Douglas-fir and bishop pine, while Arctostaphylos shrubs have been shown to provide a refuge for EMF taxa which colonize Douglas-fir seedlings after harvesting (Horton et al., 1999; Hagerman et al., 2001).
Field seedling EMF communities were dominated by Rhizopogon spp., which occurred on over half of all mycorrhizal root tips. This appeared to be partly due to the ability of the nursery taxon R. rudus to persist and colonize new roots on seedlings after planting. This was not surprising, since Castellano & Trappe (1985) found that inoculated Rhizopogon vinicolor persisted and colonized new roots on Douglas-fir seedlings after outplanting in clearcuts. Members of the genus Rhizopogon appear to establish well-distributed and persistent spore banks (Kjoller & Bruns, 2003), even in mature forests where they are not common on roots (Taylor & Bruns, 1999). Like Taylor & Bruns (1999), we found Rhizopogon spp. to be absent from mature tree roots but abundant on glasshouse bioassay seedlings.
For the genus as a whole, Rhizopogon spp. were present on 63% of root tips for seedlings > 16 m from mature trees but only on 43% for seedlings < 6 m from trees. The dominance of Rhizopogon spp. was responsible for the significantly higher proportion of taxa forming rhizomorphs on field seedlings than on nursery and glasshouse seedlings and on mature roots. The long rhizomorphs produced by Rhizopogon are thought to be adapted for long-distance transport (Agerer, 2001), but may also play a role in competing with other EMF taxa for the colonization of new roots. Simard et al. (1997) found that R. vinicolor was 20 times more abundant on untrenched Douglas-fir seedlings than on trenched seedlings, and suggested that R. vinicolor benefited from hyphal linkages to mature trees. This did not appear to be the case in our study, since seedlings > 16 m from trees had higher abundance of Rhizopogon spp. than seedlings < 6 m from trees.
Several of the EMF taxa found more frequently on seedlings < 6 m from trees have been reported to be favored by high soil nitrogen conditions, including T. sublilacina and Tylospora fibrillosa (Taylor et al., 2000; Lilleskov et al., 2002), and Russula spp. (Avis et al., 2003). Nevertheless, it seems unlikely that nitrogen would be more available near trees. In Douglas fir forests near our study sites, Barg & Edmonds (1999) found no differences in soil nitrogen dynamics at 6 m compared with 1 m from isolated mature trees. Nitrogen is often more available in the first few years after forest harvesting. Parsons et al. (1994) detected increases in soil nitrate in experimentally created gaps of 15 or more trees in lodgepole pine forests. Also, only some of the EMF taxa occurring more frequently < 6 m from trees were associated with high nitrogen substrates. For example, T. stuposa was reported by Tedersoo et al. (2003) to predominate in coarse woody debris, which usually has a high carbon to nitrogen ratio.
Ordination of EMF communities
The EMF communities of seedlings < 6 m from and > 16 m from trees were not entirely segregated when all EMF communities were included in the ordination. Nevertheless, EMF communities of seedlings < 6 m from trees had generally lower scores on the primary axis, closer to the scores for mature tree roots (Fig. 5a). The secondary axis appeared to be associated with site differences for all field samples. The EMF of mature trees had the greatest separation by site along axis 2, with low scores for Imagine, high scores for Green River and intermediate scores for Beatles. The EMF of seedlings < 6 m from mature trees followed the same pattern, while the EMF of seedlings > 16 m from mature trees had the relative position of Beatles and Imagine reversed. Glasshouse seedling EMF clustered tightly together, indicating that the site of origin of the soil in which the seedling was planted had relatively little effect on the EMF community.
Ordination of only the field seedling EMF communities revealed an effect of proximity to mature trees which was related to the primary axis. When proximity was coded as a quantitative factor (< 6 m = 1, > 16 m = 0), the treatment effect accounted for 25.5% of variation along axis 1 (r = 0.505). The secondary axis again appeared to be related to site. Imagine was associated with high values, Green River was intermediate (but highly variable) and Beatles was associated with low values along axis 2. It was notable that samples from seedlings < 6 m from trees at Green River assorted closely with samples from seedlings > 16 m from trees at the same site. This was responsible for most of the overlap between EMF communities of the two groups of seedlings. In general, differences among sites were greater than differences between seedlings < 6 m compared with > 16 m from mature trees at each site.
Ordination revealed a pattern of association between the < 6 m seedling EMF community and the mature tree community. The EMF community for seedlings > 16 m from trees was more closely associated with communities of glasshouse and nursery seedlings. It is not clear whether this similarity is a result of the direct influence of mature trees on colonization of seedlings, or a secondary effect of trees on the soil environment which facilitates the colonization or persistence of taxa associated with mature forest soils.