Lophocoleaceae

Patzak, Simon D. F., Renner, Matt A. M., Schäfer-Verwimp, Alfons, Feldberg, Kathrin, Heslewood, Margaret M., Peralta, Denilson F., de Souza, Aline Matos, Schneider, Harald & Heinrichs, Jochen, 2016, A phylogeny of Lophocoleaceae-Plagiochilaceae-Brevianthaceae and a revised classification of Plagiochilaceae, Organisms Diversity & Evolution (New York, N. Y.) 16 (3), pp. 481-495 : 489-490

Lophocoleaceae

The classification of Lophocoleaceae is in constant flux (e.g. Engel and Schuster 1984; Hentschel et al. 2007; Engel et al. 2010; Söderström et al. 2013) and far from being satisfactorily solved. Accordingly, we depict two recently used classification systems on our phylogenetic hypothesis ( Fig. 2 ) and consider a more definite treatment premature with regard to the available evidence. Main subject of controversy is the circumscription of Chiloscyphus that was often treated as a speciose cosmopolitan genus (e.g. Engel and Schuster 1984; Hässel de Menéndez 1996; Hentschel et al. 2006b) but recently reduced to two Holarctic species ( Söderström et al. 2013). The latter treatment aims to maintain several genera that were shown to nest in Chiloscyphus ( Clasmatocolea , Lophocolea , Pachyglossa, Hentschel et al. 2007 ). It is also connected with a transfer of Chiloscyphus subg. Connati to a new genus Cryptolophocolea ( Söderström et al. 2013) . Hentschel et al. (2007) resolved members of C. subg. Connati as sister to a clade consisting of Chiloscyphus s.l. and Leptoscyphus based on nrITS sequence data whereas analyses of an rbc L dataset supported a monophyletic Chiloscyphus ( Hentschel et al. 2006a) . Hentschel et al. (2007 a) thus abstained from a definite treatment of C. subg. Connati and pointed to the need of an extension of the taxon and marker sampling. Now that these data are available, C. subg. Connati / Cryptolophocolea are resolved monophyletic with strong support and sister to the remainder of Chiloscyphus s.l. with moderate support.

Aiming at monophyletic genus concepts, Söderström et al. (2013) also transferred members of C. subg. Notholophocolea to Pachyglossa although this sister relationship lacks statistical support ( Hentschel et al. 2007, this study). This transfer blurs the morphological circumscription of Pachyglossa which was earlier characterized by its polystratose leaves ( Herzog and Grolle 1958) and is also in some conflict with the acceptance of a polyphyletic Clasmatocolea ( Söderström et al. 2013) .

The generitype of Chiloscyphus , C. polyanthos , and its diploid sister C. pallescens ( Hentschel et al. 2006b) form a fully supported monophylum isolated on a relatively long branch from other species attributed to this genus. However, Chiloscyphus subg. Chiloscyphus is resolved on a polytomy alongside lineages containing species of Clasmatocolea and Pachyglossa . Even derivation of monophyletic subgeneric taxa is problematic, because sister relationships between members of this polytomy and other lineages of Clasmatocolea are unsupported. More comprehensive sampling of species from these genera may resolve Chiloscyphus subg. Chiloscyphus nested within a Clasmatocolea lineage that is also rendered paraphyletic by Pachyglossa . Maintenance of Chiloscyphus subg. Chiloscyphus in its current circumscription will probably involve proposal of new subgenera to achieve a subgeneric classification comprising only monophyla, but the number of new subgenera will not be known until all species have been included and the relationships of subg. Chiloscyphus fully resolved. These results suggest that the clarity of relationships presented by some earlier studies was an artefact of undersampling a lineage that appears to have undergone rapid early radiation. Even the more extensive species sampling and phylogeny reconstruction achieved in this study are insufficient grounds for informed decisionmaking, and more work in this part of the phylogeny is required.

With regard to the separation of Lophocolea from Chiloscyphus , we consider genera to represent a necessary evil of the binary nomenclature introduced by Linnaeus (1753); however, genera should be identifiable by morphological characters in order to be encodable in morphology-based floras. Engel and Schuster (1984) already pointed to a vast morphological overlap of Chiloscyphus and Lophocolea and thus combined them under one genus, Chiloscyphus . It is commonly seen that molecular phylogenies identify putative satellite genera as ingroup elements of larger genera and that phylogeny-based approaches lead to wider genus concepts ( Humphreys and Linder 2009). Current sampling is still inadequate to provide reliable character state reconstructions and decide on the usefulness of the deviating classification systems. A comprehensive extension of the taxon set including not only representatives of Chiloscyphus s.l. but also of the remaining lophocolean genera that have not been sampled: Deceptifrons J.J. Engel & Váňa , Evansianthus R.M. Schust. & J.J. Engel , Hepatostolonophora J.J. Engel & R.M. Schust. , Lamellocolea J.J. Engel , Leptoscyphopsis R.M. Schust. , Otoscyphus J.J. Engel, Bardat & Thouvenot , Perdusenia Hässel , Pigafettoa C. Massal. , Platycaulis R.M. Schust. , Stolonivector J.J. Engel and Xenocephalozia R.M. Schust. is necessary to arrive at a natural subdivision of this speciose liverwort lineage and to identify the morphological characters that support the clades.

Thanks to studies of Vanderpoorten and Long (2006), Devos and Vanderpoorten (2009) and Vanderpoorten et al. (2010, 2012), the phylogeny and morphological evolution of Leptoscyphus is already well understood. Here again Leptoscyphus is resolved monophyletic with weak support, though Leptoscyphus excluding subg. Austroleptoscyphus Vanderp., Schäf.-Verw. & D.G. Long is monophyletic with full support. Extension of the sampling leads to the recognition of L. gradsteinii new to Ecuador and points to the need for a recircumscription of Leptoscyphus hexagonus (Nees) Grolle. The recently proposed transfer of Lophocolea trapezoïdes Mont. to Leptoscyphus ( Söderström et al. 2013) is also confirmed. Our phylogeny points to a deep split between the Australasian L. australis and South American/ Holarctic or African species that deserves further investigation with inclusion of additional Australasian species such as Leptoscyphus compactus (Colenso) J.J. Engel and Leptoscyphus normalis (Steph.) J.J. Engel ( Engel 2015a) in future phylogenetic studies. Heteroscyphus is paraphyletic but taxon sampling still inadequate to arrive at definite conclusions. According to Söderström et al. 2016), Heteroscyphus includes some 100 species, of which we could include only 15. We thus feel also unable to comment on the subgeneric classification.

Species diversity is more evenly distributed among lineages within the Lophocoleaceae than in the Plagiochilaceae , most lineages contain several tens of species, and the two largest around 100 to 150. Total diversity within the Lophocoleaceae is approximately one third that of the Plagiochilaceae ( Söderström et al. 2016) .

Figures

Fig. 2 Maximum likelihood phylogeny of Plagiochilaceae-Lophocoleaceae-Brevianthaceae; includes Brevianthaceae and Lophocoleaceae. Deviating classification systems of Chiloscyphus are indicated