Ribosomal DNA sequences were assessed for their usefulness in distinguishing among Trichoderma isolates and for their robustness in resolving their phylogenetic relationships. DNA sequences from the D2 region of the 28S rRNA gene were determined for 50 Trichoderma isolates representing seven species. Eight distinct sequence types existed, and were mostly consistent with groupings based on morphology. Sequence variability within the D2 region alone was not sufficient to provide a reliable phylogeny. Sequences from the ITS1, 5.8S and ITS2 regions were subsequently determined for 18 of the isolates. Eight distinct ITS sequence types were detected among these 18 isolates. The ITS sequence types were generally consistent with morphology, ITS1 sequence data supported the identification of the Th3 T. harzianum group of Muthumeenakshi et al. (1994) as T. atroviride. The data also confirmed that the biocontrol strains of this study were different from those causing disease problems in the mushroom industry in Europe and North America. Results from the phylogenetic analysis stress the importance of testing the robustness of data used to predict phylogenies. Two ITS sequence data sets for the same group of isolates produced significantly different phylogenies. Congruence analysis detected that T. inhamatum (GJS90-90) was corrupting tree topologies and ‘first order pruning’ was performed by removing its sequence from the two ITS data sets. Subsequent differences in the topologies of pruned ITS1 and ITS2 trees were attributed to a lack of phylogenetic information in the ITS2 sequence region. Although ITS sequences successfully differentiated among morphologically distinct isolates within Trichoderma, it did not provide a sufficient phylogenetic signal to resolve all of their relationships.