Estimates of fungal biodiversity from environmental samples are all subject to bias. Major issues are that the commonly adopted cultivation-based approaches are suitable for taxa which grow readily under laboratory conditions, while the DNA-based approaches provide more reliable estimates, but do not indicate whether taxa are metabolically active. In this study, we have evaluated these approaches to estimate the fungal diversity in soil and freshwater samples from a subtropical forest, and compared these to RNA-based culture-independent approach intended to indicate the metabolically active fungal assemblage. In both soil and freshwater samples, the dominant taxon recovered by all three approaches was the same (Anguillospora furtiva). This taxon was cultivable from all samples and comprised 85-86 % DNA libraries and 90-91 % RNA libraries. The remaining taxa were phylogenetically diverse and spanned the Ascomycota, Basidiomycota, and Fungi incertae sedis. Their recovery was not consistent among the three approaches used and suggests that less abundant members of the assemblage may be subjected to greater bias when diversity estimates employ a single approach.