TA cloning methods are widely used in analyses of environmental microbial diversity, yet the potential of TA methods to yield phylogenetically biased results has received little attention. To test for a TA bias, we constructed clone libraries of fungal amplicons spanning the ribosomal internally transcribed spacer (ITS) and partial large subunit (LSU) from 92 boreal forest soil DNA extracts using two contrasting methods: the Invitrogen TOPO-TA system and the Lucigen PCR-SMART system. The Lucigen system utilizes blunt-ended rather than TA cloning and transcription terminators to reduce biases due to toxicity of expressed inserts. We analysed 588 clone sequences from the two libraries. Species diversity estimators applied to operational taxonomical units (OTUs) were slightly higher for Invitrogen than Lucigen, but confidence intervals for accumulation curves overlapped. Abundances of OTUs were correlated between the libraries (r(2) = 0.5, P < 0.0001), but certain OTUs had contrasting abundances in the two libraries and a likelihood ratio test rejected homogeneity of the OTU counts. We constructed parsimony and Bayesian trees from aligned LSU regions, and the 'phylogenetic test' revealed that lineage representation was not significantly different between the two libraries. We conclude that characterization of this fungal community was fairly robust to cloning method and no biases due to TA cloning were found.