Ann. Bot. Fennici 43: 349–362 ISSN 0003-3847 Helsinki 24 October 2006 © Finnish Zoological and Botanical Publishing Board 2006 Phylogeny of Vetaformaceae, Lepicoleaceae and Herbertaceae (including Mastigophoraceae) inferred from chloroplast trnL-F, nuclear ITS2, and morphology Aino Juslén Botanical Museum, Finnish Museum of Natural History, P.O. Box 7 and Plant Biology, P.O. Box 65, FI-00014 University of Helsinki, Finland (e-mail: aino.juslen@helsinki.fi) Received 31 Jan. 2006, revised version received 20 June 2006, accepted 22 June 2006 Juslén, A. 2006: Phylogeny of Vetaformaceae, Lepicoleaceae and Herbertaceae (including Mas- tigophoraceae) inferred from chloroplast trnL-F, nuclear ITS2, and morphology. — Ann. Bot. Fen- nici 43: 349–362. Phylogenetic relationships of Vetaformaceae, Lepicoleaceae, and Herbertaceae (includ- ing Mastigophoraceae) were reconstructed using chloroplast region trnL-F, nuclear ITS2, and 27 morphological characters. Forty-five species were included in the analy- sis, of which 37 belong to the ingroup. The data sets were analyzed simultaneously with direct optimization, as implemented in the program POY. The results confirm the sister relationships of Vetaformaceae and Lepicoleaceae as well as of Herbertaceae and former Mastigophoraceae. Within Lepicolea the species are divided into two sister clades. Herbertus runcinatus is sister to the rest of the genus. A clade of H. oldfieldi- anus, South American species and a species from the Azores form a separate lineage. The rest of the Herbertus species are grouped together but many unresolved nodes remain. Key words: Herbertaceae, ITS2, Lepicoleaceae, Mastigophoraceae, phylogeny, sys- tematics, trnL-F, Vetaformaceae Introduction to assume the families as the ingroup. Contrary to earlier classifications based on morphology The recent comprehensive molecular phylog- (Grolle 1983, Schuster 1984, 2000, Crandall- enies of liverworts by He-Nygrén et al. (2004, Stotler & Stotler 2000), the recent analyses that 2006) and Heinrichs et al. (2005) have resolved included sequence level data have resolved the the families Vetaformaceae, Lepicoleaceae, Mas- four families as the most derived lineage of tigophoraceae, and Herbertaceae as a mono- leafy liverworts (He-Nygrén et al. 2004, 2006, phyletic group. The group shares the possession Heinrichs et al. 2005). According to these stud- of antheridia enclosed by male bracteoles. Fur- ies, Herbertaceae and Mastigophoraceae are thermore, the liverworts in the group are mostly most closely related to some lepicoleoid species, isophyllous. As several studies using different with the Vetaformaceae as sister group to Lepi- datasets have produced the same topology and coleaceae, and Mastigophoraceae sister to Her- there is support from morphology, it is justified bertaceae. However, the sampling of the families 350 Juslén • ANN. BOT. FENNICI Vol. 43 Vetaformaceae, Lepicoleaceae, Mastigophora- ter (1984) placed Mastigophoraceae in suborder ceae, and Herbertaceae in both studies referred Ptilidiineae, which he considered to be derived. to above is restricted. This study aimed to test Crandall-Stotler and Stotler (2000) classified the presented hypothesis with more extensive Mastigophoraceae in the primitive Ptilidiineae sampling of these families. (Lepicoleales), and Vetaformaceae and Lepi- Vetaforma dusenii, the only species of Veta- coleaceae in the closely related Lepicoleiineae formaceae, has been found in Argentina and (Lepicoleales), but Herbertaceae were in a more Chile. The monogeneric Lepicoleaceae occur derived sub-order Herbertineae (Jungermanni- mostly in the southern hemisphere. Many species ales). He-Nygrén et al. (2006) placed all the have been recorded for South America and Aus- families in their Lophocoleineae in Jungerman- tralasia, but only one occurs in Africa (Schuster niales. 2000). Lepicolea was revised by Scott (1960), The relationships of Mastigophoraceae have who reported seven species. Currently at least 10 been especially controversial. Inoue (1978) sup- species are recognized (Schuster 2000). ported Grolle’s (1972) view of Mastigophora In Mastigophoraceae two genera have been belonging to Lepicoleaceae subfamily Mastigo- recognized. Dendromastigophora is a New Zea- phoroidea based on his observations of the sporo- land endemic. The genus Mastigophora has a phyte characters, which were formerly unknown. wide, disjunct range and is found in British However, Schuster (1987) rejected that place- Columbia, mid-Atlantic islands, and also from ment and reiterated his earlier idea (Schuster Oceanic Europe to the Himalayas, China and 1972, 1979, 1984) of classifying Mastigophora- to Australasia (Schuster 1987). Herbertaceae ceae in Ptilidiineae, justifying this by virtue of have been divided into two genera, Triandro- the similarities of Mastigophoraceae and Ptilid- phyllum and Herbertus. Triandrophyllum is a iaceae in their leaf division, incubous orientation small southern temperate genus (Gradstein & of leaves, and anisophylly. He also presented a da Costa 2003), while Herbertus has a circum- detailed discussion on the morphological differ- pacific range with extensions to the Appalachian ences of Lepicoleaceae and Mastigophoraceae mountains, tropical Africa and Europe (Schuster (Schuster 1987). In the classification by Cran- 2000). The taxonomy of the genera remains dall-Stotler and Stotler (2000) Mastigophoraceae mostly unresolved. remained in the suborder Ptilidiineae, which Mastigophora consists of three species. was, however, considered as a primitive subor- Monotypic Dendromastigophora was separated der among leafy liverworts. Recently, Heinrichs from Mastigophora by Schuster (1987). Her- et al. (2005) proposed to include Mastigophora- bertus is the largest genus of the group under ceae in the family Herbertaceae, based on chlo- consideration. The number of species in Her- roplast rbcL data. He-Nygrén et al. (2006) con- bertus remains uncertain; Gradstein (2001) gave firmed this result in their classification, which an estimate of ca. 25 species worldwide. Trian- was based on phylogenetic analysis of numerous drophyllum consists of 4–5 species of uncertain genes and morphology. delimitation (Schuster 2000). Gradstein (2001) In this study I address the following ques- reduced Olgantha, a third genus recognized in tions: (1) Are the families Vetaformaceae, Lep- Herbertaceae by Schuster (1996, 2000), to syn- icoleaceae, and Herbertaceae, including Mas- onymy under Triandrophyllum. tigophoraceae, monophyletic, and what are their The taxonomic position of the four fami- phylogenetic relationships? (2) Are the genera lies studied vary in recent classifications. Grolle included in this study monophyletic? (3) Can (1983) placed all four in the order Jungerman- any infrageneric relationships within the two niales. In Schuster’s (1984, 2000) classifica- larger genera, Herbertus and Lepicolea be pro- tion scheme Vetaformaceae, Lepicoleaceae, and posed? (4) Are there any evolutionary trends in Herbertaceae follow Haplomitrium, which he morphological characters within the group? In considered to be the most primitive extant liver- order to obtain answers to these questions the wort. They are all isophyllous, which Schuster trnL-F region of the chloroplast and ITS2 region (1984) believed to be a primitive feature. Schus- of the nuclear genome were sequenced, and 27 ANN. BOT. FENNICI Vol. 43 • Phylogeny of Vetaformaceae, Lepicoleaceae and Herbertaceae 351 morphological characters examined and included available in the Botanical Museum (H) of the in a combined cladistic analysis (Kluge 1989, Finnish Museum of Natural History, and also Nixon & Carpenter 1996). loaned material from various herbaria. In many cases, however, no specimen contained sexual organs, and the coding was done according to lit- Material and methods erature (Schuster 1959, 1966, 1987, 2000, 2002, Scott 1960). The characters of the matrix are Selection of terminals often used in familial or generic delimitations of the ingroup. However, no a priori assumptions Altogether 45 exemplars were studied, 37 of informative characters were made and all the belonging to the ingroup and eight to the out- characters were coded as unordered. No spo- group (Table 1). The authors for the included rophyte characters were included, because for species are given in Table 1. Representatives most of the taxa the sporophytes are unknown. of all species of the ingroup genera for which I Quantitative characters were generally excluded, had recently collected specimens available were except for those showing clear discontinuities included. For Herbertus aduncus and H. dicra- (characters 11, 14, and 18). The morphological nus, two species with wide ranges and large matrix is given in Appendix 2. morphological variation, different subspecies or samples from different geographic regions were included. Molecular data According to the results of He-Nygrén et al. (2004, 2006) the Trichocoleaceae, Plagiochi- Two mostly non-coding DNA sequence regions laceae, and Geocalycaceae form a sister group to were used for the phylogeny reconstruction, the ingroup families. I included representatives chloroplast trnL-F and nuclear ITS2. The prim- (Trichocolea tomentella, Plagiochila asplenio- ers used for sequencing the trnL-F region were ides, Chiloscyphus profundus) of each of the C and F (Taberlet et al. 1991). For ITS2 region abovementioned families in the outgroup. Addi- primers 5.SR http://www.biology.duke.edu/fungi/ tional outgroup species Blepharostoma tricho- mycolab/primers.htm and LC4-R