No abstract

In recent years, the use of extensive molecular and morphological datasets has clarified the phylogenetic relationships among the orders of complex thalloid liverworts (Marchantiidae). However, previous studies excluded extinct taxa; thereby, undersampling the actual taxonomic diversity of the group. Here, we conducted a total-evidence analysis of Marchantiidae incorporating fossils. The combined dataset consisted of 11 genes-sampled from the nuclear, mitochondrial and plastid genomes-and 128 morphological characters. Sixty-two species, representing all classes and orders within Marchantiophyta and genera within Marchantiidae were included in the analyses. Six fossils were scored from literature: two assigned to the outgroup (Metzgeriothallus sharonae and Pallaviciniites sandaolingensis) and four to the ingroup (Marchantites cyathodoides, M. huolinhensis, Ricciopsis ferganica and R. sandaolingensis). Tree searches were conducted using parsimony as the optimality criterion. Clade sensitivity was assessed across a wide range of weighting regimes. Also, we evaluated the influence of fossils on the inferred topologies and branch support. Our results were congruent with previously inferred clades above the order level: Neohodgsoniales was sister to a clade formed by Sphaerocarpales and Marchantiales. However, relationships among families within Marchantiales contradicted recent studies. For instance, a clade consisting of Monosoleniaceae, Wiesnerellaceae and Targioniaceae was sister to the morphologically simple taxa instead of being nested within them as in previous studies. Novel synapomorphies were found for several clades within Marchantiales. Outgroup fossils were more influential than Marchantiidae fossils on overall topologies and branch support values. Except for a single weighting scheme, sampling continuous characters and down-weighting characters improved fossil stability. Ultimately, our results challenge the widespread notion that bryophyte fossils are problematic for phylogenetic inference.

Our results support the groups defined by traditional taxonomy, resolve Stenokoleales nested among the lignophytes, and indicate that seed plants may share a closer ancestor with Stenokoleales than with aneurophytes. Additionally, our trees suggest a Givetian minimum age for the seed plant ancestor, a late Emsian minimum age for the Stenokoleales, and early Emsian minimum ages for lignophytes, the bilateral clade, and the aneurophyte ancestor.

enriches the documented moss diversity of an already-diverse Early Cretaceous plant fossil assemblage. This is the third moss described from the Apple Bay plant fossil assemblage and represents the first occurrence of gemma cups in a fossil moss. It is also the oldest unequivocal record of Polytrichaceae, providing a hard minimum age for the group of 136 million years.

Our results suggest that morphology is useful in resolving phylogenetic relationships in the Polytrichaceae, if both discrete and continuous characters are used. However, our rooting experiments demonstrate that there is no superior way to root analyses and indicate that relationships within the family are best evaluated using unrooted networks without outgroup taxa. These rooting problems suggest that additional information is needed to understand the phylogenetic relationships of Polytrichaceae. Such additional information could come from fossils of stem group polytrichaceous mosses, which await discovery.