Paper
Abstract Background and Aims Plant–fungus symbioses may experience temporal turnover during the ontogenetic or phenological development of the host, which can influence the ecological requirements of the host plant. In this study, we investigate temporal turnover of Ceratobasidiaceae orchid mycorrhizal fungal (OMF) communities in Prasophyllum (Orchidaceae), asking whether OMF communities are subject to temporal change attributable to orchid phenology or ontogeny. Methods Roots of adult Prasophyllum frenchii, Prasophyllum lindleyanum and Prasophyllum sp. aff. validum from Australia were sampled between autumn and spring. Seed was sown in situ as ‘baits’ to explore the mycorrhizal associations of germinating protocorms, which were compared with OMF in roots of co-occurring adult plants. Culture-dependent and -independent sequencing methods were used to amplify the internal transcribed spacer and mitochondrial large subunit loci, with sequences assigned to operational taxonomic units (OTUs) in phylogenetic analyses. Germination trials were used to determine whether fungal OTUs were mycorrhizal. Key Results A persistent core of OMF was associated with Prasophyllum, with Ceratobasidiaceae OMF dominant in all three species. Phenological turnover occurred in P. lindleyanum and P. sp. aff. validum, but not in P. frenchii, which displayed specificity to a single OTU. Ontogenetic turnover occurred in all species. However, phenological and ontogenetic turnover was typically driven by the presence or absence of infrequently detected OTUs in populations that otherwise displayed specificity to one or two dominant OTUs. Ex situ germination trials showed that 13 of 14 tested OTUs supported seed germination in their host orchid, including eight OTUs that were not found in protocorms in situ. Conclusions An understanding of OMF turnover can have practical importance for the conservation of threatened orchids and their mycorrhizal partners. However, frameworks for classifying OMF turnover should focus on OTUs important to the life cycle of the host plant, which we suggest are likely to be those that are frequently detected or functionally significant.