| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
125 Ackert Hall, Division of Biology, Kansas State University, Manhattan, Kansas 66506
A.W. Claridge
New South Wales National Parks and Wildlife Service, Threatened Species Unit, Southern Directorate, PO Box 2115, Queanbeyan, NSW 2620, Australia
J.M. Trappe
Department of Forest Science, Oregon State University, Corvallis, Oregon 97331-5752
T. Lebel
Royal Botanic Gardens, Birdwood Avenue, South Yarra, VIC 3141, Australia
D.L. Claridge
School of Resources, Environment and Society, The Australian National University, Canberra ACT 0200, Australia
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
Association analyses by contingency tables and generalized linear modeling were compared to infer relationships among hypogeous (belowground-fruiting) ectomycorrhizal fungi and potential host tree species from 136 study plots in forested habitats in southeastern mainland Australia. Results from both types of statistical approaches were highly congruent. As with previous experimental studies, no exclusive fungus-host tree associations were identified. However, the likelihood of occurrence of some species of fungi increased significantly in the presence of particular host tree species, suggesting fungal host preference or shared habitat preferences. Similarly, while most associations among fungal species were nonsignificant, a few taxa were more likely to be found in the presence of certain others. These were termed positively associated and are thought to share common climatic and microhabitat requirements or host preferences. In contrast, other combinations of fungal species were negatively associated with one another, perhaps indicating different habitat preferences. Furthermore, the finding that some fungi occurred more frequently in the presence of certain tree species provides a starting point for selection of compatible host-fungus combinations that could be used for forest nursery and restoration applications.
Key words: Acacia, Ascomycetes, Basidiomycetes, ectomycorrhizae, Eucalyptus, fungal communities, southeastern Australia
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
). Because of their beneficial effects on plant performance and growth, ectomycorrhizal fungi have been considered pivotal in reforestation and forest restoration programs (Trappe 1977). However, selection of compatible host-fungus combinations is not always simple: Fungi inoculated on nursery seedlings often fail to persist in the root systems and are out-competed by indigenous fungi when the seedlings are planted. Furthermore, ectomycorrhizal fungi easily inoculated in nurseries do not colonize all host species (Lu et al 1998).
Inoculation of eucalypts and introduced ectomycorrhizal trees with selected ectomycorrhizal fungi has been studied in Australia (Burgess et al 1993, Chilvers 1973, Chu-Chou 1979, Dunstan et al 1998, Lu et al 1998, Malajczuk et al 1982, Thomson et al 1994), but the basic ecology, host associations and habitat preferences among Australian ectomycorrhizal fungi and their hosts remain poorly understood. The communities of ectomycorrhizal fungi and host trees in native eucalypt stands need to be identified and the habitat requirements of the component species determined. Candidate fungi then can be experimentally studied for use in reforestation and restoration programs. An understanding of the factors governing species composition of fungal communities also is needed. As a step toward meeting these needs, we have used fungal and host occurrence data from earlier modeling studies to infer potential host-fungus associations that frequently occur under natural conditions in southeastern mainland Australia. We relate these host-fungus associations to tree-tree associations and infer some ecological characteristics of selected fungi based on their occurrence under natural conditions.
Host-fungus associations or host growth promoting effects of mycorrhizal fungi often have been studied by experimental inoculation studies (Bougher et al 1990, Burgess et al 1993, Chilvers 1973, Molina 1979, Molina and Trappe 1992, Thomson et al 1994, Thomson et al 1996). However, fungi often perform differently in the laboratory than in their natural environment. For example, certain media may cause atypical colonization of roots by ectomycorrhizal fungi in vitro and result in tissue disorganization more characteristic of pathogenic or parasitic infections (Duddridge 1986, Duddridge and Read 1984). Laboratory data can be difficult to extrapolate to field conditions: Fungi forming mycorrhizae in laboratory may fail to do so in the field. Pure culture studies may indicate host ranges wider than expected from sporocarp occurrence (Grand 1968, Molina 1979, 1981) because of incompatible habitat requirements among the hosts and the fungi or an inability of some fungi to compete with others under certain field conditions. Field trials addressing host and habitat compatibility are needed to rationalize application of mycorrhizal fungi in forestation and habitat restoration programs.
Claridge et al (2000a, b) described a modeling approach to understand ecological relationships of hypogeous mycorrhizal fungi in southeastern mainland Australia. They related various habitat parameters to occurrence of several widespread species. Climatic parameters (temperature and moisture) were determined as significant explanatory variables. Additional local microhabitat parameters, such as litter depth and topography, were related to the occurrence of certain fungal taxa indicating possible habitat preferences among fungi. Those results could be used to define the realized niche space in terms of the environmental optima most frequently associated with the occurrence of a given fungus (Morin 1999).
We have used data from systematic censuses of hypogeous fruit bodies and potential host trees to infer associations among ectomycorrhizal fungi and their hosts. Although association analyses have been criticized as being correlative, thus precluding inferences on causality (Olofsson et al 1999), they can identify co-occurrence patterns in plant communities (e.g. Blundon et al 1993, Choler et al 2001, Kikvidze 1993, Kikvidze and Nakhutsrishvili 1998, Sohlberg and Bliss 1984, Whitehead 1951). These co-occurrences often relate to positive mechanisms such as facilitation in plant communities (Blundon et al 1993, Choler et al 2001, Kikvidze 1993, Kikvidze and Nakhutsrishvili 1998). Thus, association analyses are valuable starting points. They help decipher relationships in large datasets and guide design of experimental studies.
We have used data on hypogeous sporocarp occurrence from four sampling events (spring and autumn 1996 and autumn 1999 and 2001) across several environmental gradients (Claridge et al 2000b). This dataset comprised 4150 collections of hypogeous fungi representing more than 200 species. Specific goals of our analyses were to infer associations: (i) between hypogeous fungus species and host species, (ii) among the common species of hypogeous fungi, and (iii) among the common species of host trees. The three types of associations were evaluated individually and together to infer ecological relationships. The conclusions of our study were compared with those from earlier habitat modeling studies (Claridge et al 2000a) to learn more about the habitat and host relationships of those fungi. Finally, we compared results obtained from traditional contingency table analyses with the more flexible, generalized linear modeling analyses.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
. In brief, 136 sites were selected in East Gippsland, Victoria, and adjacent New South Wales, Australia, and stratified by climate, geology and topography to represent dominant habitat types in the area. Within each site, a 50 x 20 m rectangle was aligned with the long axis along a contour. Hypogeous fruit bodies were revealed by raking with a four-tined garden cultivator (Castellano et al 1989) and collected during a time-constraint sampling of 100 person-minutes. The species of potential ectomycorrhizal host trees and shrubs also were recorded for each site together with a range of other microhabitat attributes. Fresh fungal fruit bodies were described with emphasis on size range, shape, surface texture and color, bruising reactions, odor, peridium thickness and the color and general texture of the gleba. These characteristics were used to tentatively assign collections to genera in the field. Collections were dried in a portable dehydrator and stored until identified to species by use of existing published and unpublished taxonomic keys and descriptions. Many of the species and some genera are undescribed and being prepared for publication. Reference collections are available at the National Herbarium, Canberra (CANB), National Herbarium of Victoria (MEL) and Oregon State University Herbarium (OSC).
The dataset from the 136 plots included more than 220 hypogeous fungal species and 26 species of ectomycorrhizal hosts. The fungi and hosts were coded as present or absent for each field site. Twenty-three species of hypogeous fungi (TABLE I) and seven species of hosts (TABLE II) occurred on 25 or more of the 136 plots, 25 being the threshold frequency we set for our association analyses. Resultant presence/absence data were arranged into two-by-two contingency tables that provide a measure of correlation or relative closeness of association among the taxa (Dighton 1994). The independence of the cell counts was evaluated by 
2 tests for fixed margin totals (Sokal and Rohlf 1995), with the number of plots (136) fixed for both hypogeous fungi and trees. These analyses were performed in SAS program PROC FREQ (SAS 1989). When the cell counts were significantly independent as judged by the 2 tests, the direction of observed associations was inferred from comparison of the observed and expected cell counts. If the observed value for a cell for co-occurrence of two taxa was greater than the expected value for that cell, the association was considered positive. If the observed value was less than the expected value, the association was considered negative.
|
|
, Nicholls 1991, Yee and Mitchell 1991). The odds ratio is best interpreted as a chance of an event. Odds ratios of 1 (a 50:50 chance) indicate an equal probability of the two possible outcomes (presence and absence of a taxon). Similarly, an odds ratio of 10 indicates that the estimated event (p: here occurrence of a taxon) is 10 times more likely than the nonevent (1 – p: here absence of a taxon). We estimated the odds ratio of hypogeous fungus occurrence as explained by presence of a host tree or another truffle species, and host tree occurrence was explained by presence of another host tree. All GLIM analyses also were performed in SAS program PROC GENMOD (SAS 1997). To evaluate effects of explanatory variables on the response variable we also estimated the parameters with their 95% confidence intervals. |
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
, IV and V display results from analyses by contingency tables and generalized linear models. Overall, results from the two types of analyses were highly congruent. In the tests of associations between host trees and hypogeous fungi, 25 of the 161 pairwise comparisons differed between the two approaches in level of significance (TABLE III). Nineteen of these, however, agreed in the direction of the association. In the tests of associations among the truffle species, only four of 243 pairwise comparisons differed in the level of significance and only one resulted in different conclusions (TABLE IV). In the tests of associations among the host trees, five of the 21 pairwise comparisons differed between the contingency tables and generalized linear models in level of significance. Three of these four agreed in the direction of the association. In summary, the conclusions of the two different analytical methods largely agreed. One main advantage of using GLIM instead of contingency tables is their ability to estimate the magnitude of the response. For example, based on the odds ratio estimated by GLIM analyses, we can conclude that Zelleromyces daucinus occurred 3–19 times more frequently on plots with Eucalyptus obliqua than on plots without it. However, because results of the analyses by contingency tables largely agreed with those of the generalized linear regression and their interpretation is simple (Dighton 1994), we present our results according to the contingency tables. Furthermore, the contingency tables are well suited for the specific question posed in our analyses: Do two species occur together more frequently than expected by random chance?
|
|
) do not suggest specificity of any single fungus for any host; in no case did fungus species occur only on plots with a particular host species.
Many of the associations were nonsignificant (TABLE III). According to the 2 tests, the cell counts in the contingency tables were dependent and did not suggest associations deviating from what could have been obtained by random chance. For example, Castoreum radicatum, Descomyces albus, Gymnomyces eburneus, Hysterangium rhodocarpum and Zelleromyces microsporus were not significantly associated with any of the seven host species examined.
Some positive associations were significant (TABLE III, FIGS. 1, 2). These can be interpreted as indicating either or both host preferences of the fungi or similar habitat requirements among the fungi and their hosts. For example, Cortinarius globuliformis was highly significantly associated with Eucalyptus dalrympleana; both occur in cool, montane environments. According to our generalized linear model for this relationship, C. globuliformis was 1.07–5.70 times more frequent on sites with E. dalrympleana than on sites without. Hysterangium inflatum, Protoglossum levisporum and Stephanospora flava were associated positively with Eucalyptus cypellocarpa and E. muelleriana, all of which occur either in mild, well-drained or coastal environments. Cystangium seminudum, Descomyces albellus, Hydnangium carneum and Zelleromyces daucinus were highly significantly associated with Acacia dealbata. The first two showed either negative or nonsignificant associations with the eucalypt species, whereas the second two had significant positive associations with one or more Eucalyptus spp.
|
|
, FIGS. 1, 2). Acacia dealbata and Eucalyptus dalrympleana were negatively associated with Hysterangium affine, P. levisporum, and S. flava. Acacia dealbata also was associated negatively with H. inflatum and E. dalrympleana with Zelleromyces corkii. All these negatively associated fungus species (except H. affine) were positively associated with at least one of the four host species forming the positively associated host tree cluster (FIG. 2).
Although no exclusive, specific associations between the host trees and hypogeous fungi were detected, some were strong (TABLE III). Detection of Hydnoplicata convoluta, for example, was 2.18–21.98 times more likely on plots with Eucalyptus obliqua than on plots without. This odds ratio translates to 69–96% probability of detecting H. convoluta when E. obliqua was present.
Fungus x fungus associations. –
Results of most of the fungus x fungus tests were nonsignificant, but the 23 taxa did group into two distinct clusters, each characterized by multiple interconnecting positive associations (TABLE IV, FIG. 3). The two clusters broadly follow the associations between host trees and fungi (FIGS. 1, 2). The exceptions include fungal species that occur in association with hosts from both of the two host clusters (e.g. Descomyces giachinii, Z. corkii, Z. daucinus). Significant positive associations emerged in several other cases (TABLE IV, FIG. 3). The Chamonixia vittatispora complex, C. globuliformis and Hydnoplicata convoluta were highly significantly associated. Castoreum radicatum was highly significantly associated with C. globuliformis but not with the other two; C. radicatum was highly significantly associated with Hysterangium affine as well. Cortinarius globuliformis also was significantly to highly significantly associated with P. levisporum, Hysterangium rhodocarpum, Mesophellia glauca, and Zelleromyces claridgei. Cystangium megasporum was significantly to highly significantly associated with Cystangium seminudum and the four Zelleromyces spp. Noteworthy clustering occurred within genera. The three Descomyces spp. were significantly to highly significantly associated with each other as well as with H. carneum. Three of the four Zelleromyces spp. were highly significantly associated with each other (Z. claridgei was highly significantly associated with Z. daucinus but not with the other two Zelleromyces spp.).
|
Tree x tree associations. –
Our primary objective was to identify associations between fungi and hosts and among the fungi. However, analysis of associations among the tree species adds interpretive value to consideration of ecology of the fungi. The analyses of the seven qualifying species revealed one distinct community type and a group of four species linked in various combinations (TABLE V, FIG. 4). Acacia dealbata, Eucalyptus dalrympleana and E. radiata were highly significantly associated with each other. In contrast, these three either were significantly negatively associated or showed no significant associations with the other four species that grouped together. Among the other four species, E. cypellocarpa was significantly to highly significantly associated with both E. muelleriana and E. obliqua. In addition to E. cypellocarpa, E. muelleriana was significantly associated with E. sieberi but showed a significant negative association with E. obliqua.
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
. Choler et al (2001) noted that association analyses, although criticized by some as being only correlative, prove valuable in determining occurrence of positive interactions in plant communities. The results of our analyses provide an understanding of where individual species and communities of species occur. This understanding is a starting point for selection of compatible fungus-host combinations for forest nursery and restoration applications. Fungus x tree associations. – Although some associations between fungi and trees were significant and positive, none of the 23 truffle species occurred exclusively in association with a single host subgenus or species, indicating no strict host specificity. All seven potential host species, in turn, were associated positively with more than one species of hypogeous fungus. The common hypogeous fungus species included here clearly have broad host ranges but share habitat preferences similar to the hosts they are positively associated with. Perhaps the rarer species in the study are the ones to examine for host specificity, but their scarcity precluded inclusion in the association analyses conducted here.
Earlier studies have failed to detect any ectomycorrhizal fungus specificity to particular eucalypt species (Chilvers 1973, Malajczuk et al 1982), although some host-fungus combinations prove incompatible and form no mycorrhizas in synthesis assays (Lu et al 1998, Malajczuk et al 1982). Our analyses found no host specificity, even though the trees included in our study represent two families (Mimosaceae, Myrtaceae), two genera (Acacia, Eucalyptus), and within Eucalyptus two subgenera and five series. Host specificity of ectomycorrhizal fungi occurs at the genus or rarely at the subgenus level of host trees (Molina et al 1992), so no specificity was expected below the subgenus level in Eucalyptus. However, the analyses enable inferences beyond simple fungus-host associations to include shared habitats as well. Data on fungal co-occurrences can be used to identify similarities in habitat requirements among ectomycorrhizal fungi. As a result, from the presence of fruit bodies of one species these data predict that certain others are likely to be present or at least be amenable to that habitat.
Most of the observed fungus x tree associations were nonsignificant, indicating that factors other than any of the seven host species determined occurrence. Claridge et al (2000a) determined environmental parameters that defined habitat preferences for some common hypogeous fungi in southeastern Australia. At a landscape level, climatic factors, including temperature and moisture, best explained occurrence of some fungal species. Additional micro-habitat factors such as disturbance history and soil fertility were correlated with fungal occurrence at a more local scale. These factors likely control the occurrence and distribution of both fungi and host trees.
Interpreting significant negative associations is more straightforward than nonsignificant ones. The fungi clearly differ in their environmental requirements. At each trophic level they manifest a range of environmental conditions under which they occur, thereby defining their niche. The negative associations can be inferred from dissimilar habitat preferences. The negative associations among the hypogeous fungi and their ectomycorrhizal hosts merely indicate that the species occur together less frequently than expected by random chance.
Fungus x fungus associations. – A major conclusion to be drawn from the analyses of associations among fungal species is that most taxa have evolved individual habitat requirements that thwart definition of consistent fungal communities by association analysis. Nonetheless, the highly significant positive association of the C. vittatispora species complex with C. globuliformis was interesting because it confirmed an inference from our field observations. Cortinarius globuliformis forms distinctive, bright yellow hyphal mats in the upper soil layers. Chamonixia fruit bodies were observed within these mats, often side by side with fruit bodies of Cortinarius. Although no obvious a side-by-side association was noted with Hydnoplicata convoluta, that species was collected commonly at the same sites and highly significantly associated with the other two species.
Three Descomyces species were associated highly significantly with H. carneum. These four species occur over a great diversity of sites. Those species also commonly appear in disturbed sites, such as within re-growth on clearcuts, road verges or plantations of eucalypts abroad (Trappe and Claridge unpubl data, Giachini et al 2001). This generally nonspecific habitat adaptability, as inferred by lack of significant association with other fungal species, may be hypothesized from our data for G. westresii, which show a highly significant association with two of the three Descomyces species but otherwise is not associated with other fungi. Fungi with broad habitat adaptability might be good candidates for inoculation of seedlings in container nurseries, especially in ecological restoration programs.
The identification of positive associations among the fungal species also highlight the importance of factors other than the host tree species. Cortinarius globuliformis, for example, though associated positively only with E. dalrympleana, was associated positively with other fungi that in turn were associated positively with other eucalypt species. Again, nonsignificant, negative and positive associations were observed. The positive associations among the species of fungi suggest either similar host and/or habitat preferences, whereas negative associations infer differences in these preferences.
Tree x tree associations. –
Although the associations among trees were analyzed particularly to aid in interpretation of fungal associations, the grouping of the seven tree species into two highly significantly negatively associated clusters was unexpected. A close habitat/tree community relationship was revealed between A. dealbata and E. dalrympleana and E. radiata. Acacia dealbata is distributed over a wide range of elevations and moisture regimes in southeastern Australia, whereas the two eucalypts characteristically inhabit cool, moist and generally montane habitats (Boland et al 1984). The other eucalypts in our analyses mostly inhabit warmer, low-elevation habitats with well drained soils (Boland et al 1984). These data suggest that A. dealbata is more habitat specific in our study area than its widespread distribution would have led us to presume. Clearly its habitat requirements involve environmental characteristics beyond inferences from simple association with other tree species.
The other four eucalypt species show a complex set of associations, but all are significantly to highly significantly negatively associated with the A. dealbata-E. dalrympleana-E. radiata cluster. The significantly negative association of E. muelleriana and E. obliqua indicates differing habitat characteristics of the two, whereas their significant to highly significant association with both E. cypellocarpa and E. sieberi suggests that the latter two species bridge the habitat ecological gap between the former two.
Inferences from integrating the three association analyses. – In general, the fungi positively associated with the A. dealbata-E. dalrympleana-E. radiata cluster either were not significantly or significantly negatively associated with the E. cypellocarpa-E. muelleriana-E. obliqua-E. sieberi group and vice versa. Of the 23 fungal species analyzed, only three bridged the two tree groups by a significant positive association with at least one tree species in each group: D. giachinii, Z. claridgei and Z. daucinus. Four species showing one or more significant positive associations with only one of the two tree species groups also showed significant negative association with one or more tree species of the second tree group: D. albellus, H. carneum, P. levisporum and S. flava.
Descomyces albellus is positively associated with A. dealbata and E. dalrympleana but negatively with E. radiata, even though those three trees are associated positively with each other. One can infer that the site characteristics that permit E. radiata to be present with A. dealbata and E. dalrympleana in some way are inimical to D. albellus; we cannot conceive of any reason why E. radiata would antagonize D. albellus. The last tree species also is associated negatively with the other two tree species, so in that case one can infer that the site characteristics favorable to E. sieberi were unfavorable for both D. albellus and the other two tree species.
Inferences from combined habitat modeling and association analyses. –
The association analyses presented here are based on four collection events spread over 6 y. The habitat modeling of individual fungal species by Claridge et al (2000a, b) was based on collections from only the first two samplings of these same plots in spring and autumn 1996, but the models incorporated environmental variables of climate, geology, topography and (among other microhabitat variables) amounts of coarse, woody debris and litter on the forest floor. The data at that time permitted satisfactory modeling of six species (see Claridge et al 2000a). The occurrence of two additional species, H. convoluta (designated as H. whitei by Claridge et al 2000a) and Hysterangium rhodocarpum could not be statistically explained by any of the environmental variables measured by Claridge et al (2000a, b).
The explanatory model for the C. vittatispora species complex (Claridge et al 2000a) indicated that the probability of its occurrence increased with relatively low temperatures during the hottest month of the year, a moderate annual mean moisture regime (i.e., neither the low or high ends of the moisture scale) and relatively shallow litter. Cortinarius globuliformis, highly significantly associated with the C. vittatispora species complex, had increased probability of occurrence with falling temperature of the coldest month of the year and less moisture (Claridge et al 2000a). Although the explanatory models for these two species produced significant probabilities in different parameters, both suggest they primarily would occupy relatively cool sites with relatively low to moderate moisture status, as represented in the study area. The association analysis showed a highly significant, positive association between the two species. However, neither the Chamonixia nor the Cortinarius were associated significantly with the same tree species. The occupation of similar sites evidently does not relate to a coincidence of particular tree species but to other factors. For example those that proved significant in the explanatory models: litter depth (Chamonixia) or years since fire (Cortinarius).
Hysterangium inflatum in the explanatory model showed increased probability of occurrence with higher temperature of the coldest month, sites in gullies and greater litter depth but decreased probability with greater amounts of coarse woody debris (Claridge et al 2000a). In the association analyses among fungi, H. inflatum was highly significantly positively associated with S. flava and both fungi shared a highly significant positive association with E. cypellocarpa, E. muelleriana and E. obliqua, as well as a significant negative association with A. dealbata. Data available when the explanatory models were developed (Claridge et al 2000a) were inadequate to model S. flava, but its highly significant association with H. inflatum and their congruent positive and negative associations with tree species suggest a strong similarity in habitat requirements.
The explanatory model for Z. microsporus showed that its probability of occurrence increased with increasing soil nitrogen and litter depth. This was the only Zelleromyces sp. that qualified for modeling at that time. The present association analyses show highly significant positive associations among Z. microsporus, Z. corkii and Z. daucinus. These three species, however, do not show similar positive associates with tree species. The habitat factors primarily accounting for their association can be hypothesized as those identified in the explanatory model for Z. microsporus (i.e., relatively high soil-nitrogen and litter depth).
Other differences between what could be inferred from the explanatory models versus the association analyses are difficult to interpret at present, except to say that positive associations with tree species do not relate unequivocally to the explanatory habitat models or associations among fungi. In part this is because each fungus species has its own array of habitat relationships, as shown by the explanatory models. For example, in explanatory models the probability of occurrence of the Chamonixia complex and P. levisporum fell with increasing litter depth, whereas that of H. inflatum and Z. microsporus rose with increasing litter depth. Litter depth can be affected by site history (i.e., time since last fire) and thus could not be expected to relate directly to tree species, geology or climate of the site.
Conclusions. –
In experiments on fungal fruit-body distribution related to thickets of Rhododendron maximum in the southern Appalachians, Walker and Miller (2002) concluded ‘‘Differences within the ectomycorrhizal fungus community were associated only weakly with environmental parameters, as indicated by ordination.’’ Their data were taken from 4 m2 experimental plots ‘‘on a single hillside with similar aspect and slope.’’ Our data, in contrast, were taken from time-constraint sampling of 136 plots of 1000 m2 representing a wide range of measured variables, including those of forest communities, soils, climate, fire history and coarse, woody debris (Claridge et al 2000b). The four sampling events conducted on these plots over 6 yr yielded 4150 collections representing more than 220 species. The data from the first two samplings (1540 collections) permitted habitat modeling of nine of the more common species, with abiotic factors prominent as explanatory factors (Claridge et al 2000a). The association analyses in the present paper, based on all 4150 collections, indicate that biotic factors also can be important, at least in that certain trees and fungi have overlapping environmental requirements, as indicated by their occurrence in similar habitats. Studies such as those of Walker and Miller (2002) combined with ours highlight that each ectomycorrhizal fungus has its own suite of habitat requirements that can be explained only by assessment of a broad range of abiotic and biotic parameters. Through integration of modeling and association analyses, our results expand knowledge and understanding of why certain fungi occur in some habitats but not in others.
Given an adequate database, explanatory models for the occurrence of ectomycorrhizal fungal species and association analyses can be combined to increase our understanding of the ecology of those fungi. Such data can be helpful in defining their habitat requirements, host preferences, successional patterns and community structure. They also can provide insights useful in restoration ecology, both for identifying the fungi with particular promise as inoculum for restoring forests in long-cleared areas and, in the case of the hypogeous species, for establishing the food base that is important for many animals (Maser et al 1984, Claridge and May 1994).
As a result of our additional sampling, the database for the hypogeous fungal species previously modeled has been strengthened substantially and the added frequency of more species will permit development of additional explanatory models. This progress will sharpen our interpretive capability for these ecologically important fungi.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
1 Corresponding author. E-mail: ari{at}ksu.edu
|
LITERATURE CITED |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
Blundon DJ, MacIsaac DA, Dale MRT. 1993. Nucleation during primary succession in the Canadian Rockies. Can J Bot 71:1093–1096.
Boland DJ, Brooker MIH, Chippendale GM, Hall N, Hyland BPM, Johnston RD, Kleinig DA, Turner JD. 1984. Forest trees of Australia. 4th ed. Collingwood, Australia: CSIRO Publishing. 687 p.
Bougher NL, Grove TS, Malajczuk N. 1990. Growth and phosphorus acquisition of karri (Eucalyptus diversicolor F. Muell.) seedlings inoculated with ectomycorrhizal fungi in relation to phosphorus supply. New Phytol 114:77–85.
Brooker MIH. 2000. A new classification of the genus Eucalyptus L’Her. (Myrtaceae). Aust Syst Bot 13:79–148.
Burgess TI, Malajczuk N, Grove TS. 1993. The ability of 16 ectomycorrhizal fungi to increase growth and phosphorus uptake of Eucalyptus globulus Labill. and E. diversicolor. F. Muell. Plant Soil 153:155–164.
Castellano MA, Trappe JM, Maser Z, Maser C. 1989. Key to spores of the genera of hypogeous fungi of north temperate forests with special reference to animal mycophagy. Eureka, California: Mad River Press. 186 p.
Chilvers GA. 1973. Host range of some eucalypt mycorrhizal fungi. Aust J Bot 21:103–111.
Choler P, Michalet R, Callaway RM. 2001. Facilitation and competition on gradients in alpine plant communities. Ecology 82:3295–3308.
Chu-Chou M. 1979. Mycorrhizal fungi of Pinus radiata in New Zealand. Soil Biol Biochem 11:557–562.
Claridge AW, Barry SC, Cork SJ, Trappe JM. 2000a. Diversity and habitat relationships of hypogeous fungi. II. Factors influencing the occurrence and number of taxa. Biodiv Conserv 9:175–199.
———, Cork SJ, Trappe JM. 2000b. Diversity and habitat relationships of hypogeous fungi. I. Study design, sampling techniques and general survey results. Biodiv Conserv 9:151–173.
———, May TW. 1994. Mycophagy by Australian mammals. Aust J Ecol 19:251–275.
Dighton J. 1994. Analysis of micromycete communities in soil: a critique of methods. Mycol Res 98:196–198.
Duddridge JA. 1986. The development and ultrastructure of ectomycorrhizas. IV. Compatible and incompatible interactions between Suillus grevillei (Klotzsch) Sing. and number of ectomycorrhizal hosts in vitro in the presence of exogenous carbohydrate. New Phytol 103:465–471.
———, Read DJ. 1984. Modification of the host-fungus interface in mycorrhizas synthesised between Suillus bovinus (Fr.) O. Kuntz and Pinus sylvestris L. New Phytol 96:583–588.
Dunstan WA, Dell B, Malajczuk N. 1998. The diversity of ectomycorrhizal fungi associated with introduced Pinus spp. in the Southern Hemisphere, with particular reference to Western Australia. Mycorrhiza 8:71–79.
Giachini A, de Olivieria VL, Trappe JM, Castellano MA. 2000. Ectomycorrhizal fungi in Eucalyptus and Pinus plantations in southern Brazil. Mycologia 92:1166–1177.
Grand LF. 1968. Conifer associates and mycorrhizal syntheses of some Pacific Northwest Suillus species. For Sci 14:304–312.
Kikvidze Z. 1993. Plant species associations in alpine-subnival vegetation patches in the Central Caucasus. J Veg Sci 4:297–302.
———, Nakhutsrishvili G. 1998. Facilitation in subnival vegetation patches. J Veg Sci 9:297–302.
Lu X, Malajczuk N, Dell B. 1998. Mycorrhiza formation and growth of Eucalyptus globulus seedlings inoculated with spores of various ectomycorrhizal fungi. Mycorrhiza 8:81–86.
Malajczuk N, Molina R, Trappe JM. 1982. Ectomycorrhiza formation in Eucalyptus. I. Pure culture synthesis, host specificity and mycorrhizal compatibility with Pinus radiata. New Phytol 91:467–482.
McCullagh P, Nelder JA. 1989. Generalized linear models. 2nd ed. London, England: Chapman & Hall. 499 p.
Molina R. 1979. Pure culture synthesis and host specificity of red alder mycorrhizae. Can J Bot 57:1223–1228.
———. 1981. Ectomycorrhizal specificity in the genus Alder. Can J Bot 59:325–334.
———, Massicotte H, Trappe JM. 1992. Specificity phenomena in mycorrhizal symbioses: community-ecological consequences and practical implications. In: Allen MF, ed. Mycorrhizal functioning—an integrative plant-fungal process. New York: Chapman & Hall. p 357–423.
———, Trappe JM. 1992. Biology of the ectomycorrhizal genus, Rhizopogon. I. Host associations, host specificity and pure culture syntheses. New Phytol 126:653–675.
Morin PJ. 1999. Community ecology. New York: Blackwell Science. 320 p.
Nicholls AO. 1991. Examples of the use of generalized linear models in analysis of survey data for conservation evaluation. In: Margules CR, Austin MP, eds. Nature conservation: cost effective biological surveys and data analysis. Canberra, Australia: CSIRO Publications. p 54–63.
Olofsson J, Moen J, Oksanen L. 1999. On the balance between positive and negative plant interactions in harsh environments. Oikos 86:539–543.
Peintner U, Bougher NL, Castellano MA, Moncalvo J-M, Moser MM, Trappe JM, Vilgalys R. 2001. Multiple origins of sequestrate fungi related to Cortinarius (Cortinariaceae). Amer J Bot 88:2168–2179.
SAS. 1997. SAS/STAT Software: Changes and enhancements through release 6. 12. North Carolina: SAS Institute Inc.
———. 1989. SAS/STAT user’s guide, version 6. 4th ed. Vol. 1. Cary, North Carolina: SAS Institute Inc.
Smith SE, Read DJ. 1997. Mycorrhizal symbiosis. 2nd ed. London, England: Academic Press. 605 p.
Sohlberg EH, Bliss LC. 1984. Microscale pattern of vascular plant distribution in two high arctic plant communities. Can J Bot 62:2033–2042.
Sokal RR, Rohlf FJ. 1995. Biometry. 3rd ed. New York: W.H. Freeman & Co. 887p.
Thomson BD, Grove TS, Malajczuk N, Hardy GESJ. 1994. The effectiveness of ectomycorrhizal fungi in increasing the growth of Eucalyptus globulus Labill. in relation to root colonization and hyphal development in soil. New Phytol 126:517–524.
Thomson BD, Hardy GESJ, Malajczuk N, Grove TS. 1996. The survival and development of inoculant ectomycorrhizal fungi on roots of outplanted Eucalyptus globulus Labill. Plant Soil 178:247–253.
Trappe JM. 1977. Selection of fungi for ectomycorrhizal inoculation in nurseries. Ann Rev Phytopathol 15:203–222.
Walker JF, Miller OK Jr. 2002. Ectomycorrhizal sporophore distributions in a southeastern Appalachian mixed hardwood/conifer forest with thickets of Rhododendron maximum. Mycologia 94:221–229.
Whitehead FH. 1951. Ecology of the Altiplano of Monta Maiella, Italy. J Ecol 86:405–420.
Yee TW, Mitchell ND. 1991. Generalized additive models in plant ecology. J Veg Sci 2:587–602.
Zdravevski J. 2000. A taxonomic and biogeographic review of the hypogeous fungal genus Chamonixia in southeastern Australia [Honor’s Thesis]. Melbourne, Australia: University of Melbourne. 100 p.
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
p < 0.01; + or –; 0.01 
