Reinstatement of Chiastocaulon carl (Plagiochilaceae), based on evidence from nuclear ribosomal ITS and chloroplast gene rps4 sequences

Morphological approaches have led to controversial opinions regarding the systematic position of the Asian Junger-mannia dendroides Nees within the family Plagiochilaceae. In recent times, the taxon was treated both as genus Chiastocaulon Carl, as a member of Plagiochila sect. Dendroideae Gottsche, Lindenb. and Nees, and as a representative of Plagiochila subgen. Chiastocaulon (Carl) Inoue. Sequences of the chloroplast-encoded rps4 gene and the internal transcribed spacers of nuclear ribosomal DNA of 28 representatives of the Plagiochilaceae and Herbertus subdentatus (Herbertaceae, outgroup) were obtained to test the different hypotheses. Maximum likelihood analyses were performed, both on the separate and combined data sets, and in all cases resulted in a single optimal topology. The different phylogenies were congruent, but the analyses of the combined data set led to an overall more significant bootstrap support. Jungermannia dendroides was placed in a moderately supported clade with Pedinophyllum interruptum and Plagiochilion mayebarae, sister to the well supported Plogiochila clade. The topology justifies the recognition of Jungermannia dendroides as genus Chiastocaulon. Plagiochila frondescens (Nees) Lindenb., P. fruticosa Mitt. and P. pulcherrima Horik., placed alongside Chiastocaulon/Plagiochila dendroides by some authors, form a robust clade within Plagiochila (Dumort.) Dumort., assignable to P. sect. Fruticosae Inoue. The three species share the dendroid habit and alternating foliation with Chiastocaulon dendroides but lack well developed ventral intercalary branches. 11 sectional clades with robust bootstrap support were identified within the Plagiochila lineage. Many morphological characters of monophyletic lineages within the Plagiochilaceae appeared homoplastic, indicating that a natural subdivision of the family is only possible by the integration of molecular data. The molecular topologies justify a hierarchical subdivision of Plagiochila at some point in the future; however, well supported sectional groups of the molecular trees will presumably lack morphological autapomorphisms.