The Lejeunea tumida species group is positively polyphyletic (Lejeuneaceae: Jungermanniopsida) Matt A. M. Renner A , Elizabeth A. Brown A,C and Glenda M. Wardle B A National Herbarium of New South Wales, Mrs Macquaries Road, Sydney, NSW 2000, Australia. B School of Biological Sciences, Heydon Laurence Building A08, University of Sydney, Sydney, NSW 2006, Australia. C Corresponding author. Email: Elizabeth.Brown@rbgsyd.nsw.gov.au Abstract. A phylogeny based on nrITS1 and trnL–F sequences resolves the Lejeunea tumida species group polyphyletic with individuals belonging in two clades either side of the basal-most node within Lejeunea. It is impossible for the Lejeunea tumida species group to be more polyphyletic and still be attributed to the same genus under the existing generic classification. A simulation-based approach to testing the null hypothesis of group monophyly rejects this at the P < 0.01 level of significance. Bayesian tests find very strong support for polyphyly, given the data. The monophyly of L. tumida s.s. + L. colensoana is fully supported; however, although Lejeunea tumida s.s. is nested within L. colensoana, this position is not supported. Both L. oracola and L. rhigophila are resolved as monophyletic. Whereas there is moderate support for the monophyly of L. rhigophila, there is no support for the monophyly of L. oracola. Neither is the monophyly of L. oracola + L. rhigophila supported in Bayesian or parsimony analysis. Introduction With a handful of exceptions (Pfeiffer et al. 2002), new liverwort species have been described because they are morphologically distinguishable, in terms of either overall similarity or specific character differences. Although morphological discontinuity is the justified basis of the taxonomic species concept (de Queiroz 1998), during the past decade, concerns have been raised about the ability of morphology to provide conclusive data on both establishing relationships between species and establishing species boundaries (Reiner-Drehwald and Goda 2000; Gradstein et al. 2003; Hienrichs et al. 2004; Ilkiu-Borges 2005). Morphology may confuse and confound endeavours to resolve relationships because of lack of stable morphological boundaries between species (Feldberg and Hienrichs 2006), and variability in gametophytic characters (Hienrichs et al. 2004). Both may result in too much being read into subtle morphological differences that do not reflect phylogenetic groups. In the Lejeuneaceae, both scenarios have been demonstrated. Hartmann et al.(2006) found that five characters that were used to circumscribe species of Bryopteris (Stotler and Crandall-Stotler 1974) were diffusely distributed, and formed no sound basis for identifying morphological groups congruent with molecular clades. In contrast, Hienrichs et al.(2009) found that despite clear genetic separation, there was extensive morphological overlap among phylogenetic entities within Marchenisia brachiata (Sw.) Schiffn. The decoupling of morphological and molecular variation and resultant conflict between morphological and molecular groups observed in the Lejeuneaceae, and a range of other bryophyte lineages (Shaw and Allen 2000; Vanderpoorten 2004), demonstrates that morphological data must be interpreted with caution in bryophytes (Hienrichs et al. 2003). A cautious approach to interpretation of morphology has generally resulted in broad morphological-species concepts being applied (Reiner-Drehwald and Goda 2000; Hienrichs et al. 2001, 2004; Zhu and Gradstein 2005; Burghardt and Gradstein 2008), resulting in synonymisation of species justified by the observation of absence of morphological discontinuity. Lejeunea tumida Mitt. was described by Mitten (1855) for plants with acute underleaf lobes, tumid lobules and inflated perianths. Mitten cited three specimens in his protologue, but did not identify a type. The first specimen listed by Mitten, collected by Sinclair near Auckland, was designated lectotype by Grolle (1982). Stephani (1896) described Taxilejeunea colensoana Steph., which also possessed tumid perianths. Both L. tumida and T. colensoana are small pellucid plants, and this aspect, in combination with their shared possession of inflated perianths led Schuster (1963a) to the conclusion that T. colensoana was ‘evidently a synonym’ of L. tumida, a view also accepted by Grolle (1982). The inflated perianths of L. tumida are highly unusual (Reiner-Drehwald and Schäfer-Verwimp 2008), and as L. tumida s.l. was the only regional species with this feature it was arguably New Zealand’s most distinctive Lejeunea species. This distinctiveness led Schuster (1963a) to propose a new monotypic subgenus for L. tumida. Lejeunea subg. Sphaerocolea R.M. Schust. was characterised by its (1) multicellular lobular apical tooth, (2) heavily inflated lobules and (3) inflated ecarinate perianths. Schuster (1963a) noted that the lobules of L. tumida s.l. were ‘quite unique within Lejeunea s.lat., the 2-celled apical tooth is found again in no other Lejeunea known to me, and the almost ovoid-spheroid lobule form is also without any exact Ó CSIRO 29 April 2011 10.1071/SB10047 1030-1887/11/010010 CSIRO PUBLISHING www.publish.csiro.au/journals/asb Australian Systematic Botany, 24, 10–18