Herpetineuron

Systematic position of Herpetineuron View in CoL

The grouping of Herpetineuron with Anomodon was suggested by Brotherus (1907, 1925), who placed these genera in Anomodontoideae (one out of four subdivisions of the Thuidiaceae ). Subsequent authors placed it in the Thuidiaceae , e.g. Watanabe (1972), Crum & Anderson (1981), Noguchi (1991). Buck & Vitt (1986) resurrected the Anomodontaceae , including Herpetineuron and some less closely related genera (discussed above). Granzow-de la Cerda (1997) provided a revision of Anomodon and Herpetineuron , but he avoided discussion of the family-level taxonomy.

Thus, no justification of the close relationship between Herpetineuron and Anomodon has ever been put forward. The molecular phylogenetic analysis (based on rbc L) of Tsubota et al. (2004) supported this position, whereas the analysis of Cox et al. (2010) (based on rps4 and nad5) did not. Not one of five studied loci supported the relationship of Herpetineuron with Anomodon ; most markers left Herpetineuron unresolved in the polytomy outside the Anomodon -clade and the Neckeraceae + Lembophyllaceae + Orthostichellaceae clade, and ITS indicated that its closest relationship was with the Thuidiaceae , with moderate support. In view of the isolated position of Herpetineuron , usually on very long branch, which reflects its strong divergence and peculiar morphology, we foresee that Herpetineuron needs its own family. However as sampling of the Thuidiaceae is not sufficient in the present study, we refrain from doing this now, and temporarily suggest returning Herpetineuron to the Thuidiaceae .

Some morphological features of Herpetineuron are remarkable. Its leaf structure is especially outstanding and merits a short comment here. The genus Herpetineron includes only a single species, H. toccoae (Sull. & Lesq.) Cardot , a widespread pantropical moss penetrating to the temperate zone in the eastern sectors of both Eurasia and North America. It grows mostly in habitats which experience moderately long periods of severe desiccation, requiring the ability to incurve its leaf margins in the dry state ( Fig. 18A, B, I View Fig ) and spread them after wetting. The peculiarity of Herpetineuron toccoae is an exceedingly conspicuous ontogenetic sector of 4 × 4, 4 × 8, and 8 × 8 cells throughout the leaf ( Fig. 18C–H View Fig ). Such sectors are universal in mosses, as the moss leaf is always formed from a single cell by a number of divisions ( Schimper, 1860b; Frey, 1970); they are usually masked by developmental irregularities, and far less apparent in fully developed leaves ( Donskov, 2015). We don’t know of any other moss whose ontogeneric sectors are so clearly apparent. Our tentative explanation of this expression is that the repeated leaf incurvation requires reliable joints, and the ontogenetic sectors in Herpetineuron are connected by efficient joints (e.g. Figs. 18L, N View Fig ), often associated with papilla-like thickening ( Figs. 18M, P View Fig ). Note that in addition to incurved leaf margins, the leaves of Herpetineuron in the dry state are also incurved along their length.

* * *

The occurrence of papillae and their structural variation within phylogenetic lineages

The present study found mosses with papillose laminal cells in two families where they were either unknown ( Lembophyllaceae ) or extremely rare ( Neckeraceae ).

It seems that the taxonomic value of papillae on the leaf lamina has been strongly overestimated in moss systematics. Some genera or species with papillose laminal cells, previously placed in various families based on this character, have already been found to have different affinities, like Myurella in the Plagiotheciaceae , while some other groups, e.g. Claopodium , actually a relative of Brachytheciaceae , are still awaiting a final verdict. The present analysis indicated that the latter genus has no affinity with either Anomodon or the Thuidiaceae , where it was traditionally placed.

Many pleurocarpous moss families include species with cells that are prominently prorate to “dorsally papillose in the distal cell end” but have not evolved further to a central-unipapillose or multipapillose state. Examples are found in the Brachytheciaceae ( Brachythecium (Bryhnia) novae-angliae ), Pseudoleskeellaceae ( Pseudoleskeella papillosa ), Amblystegiaceae ( Palustriella decipiens ), and Hylocomiaceae ( Hylocomiadelphus triquetrus ).

There are genera which, in different species, develop papillae at the distal end of the cell or in a central position ( Myurella , Haplocladium , Pseudoleskea ); note that in these cases (and maybe in most others) the papillae occur only on the dorsal leaf surface. This can lead to two trends in the complexity in the pattern of papillae, firstly from unipapillose to multipapillose (observed in Claopodium and Thuidium ), and secondly from dorsally papillose to papillose on both surfaces. The latter trend is gradual, and papillosity on the ventral surface is often less prominent compared with the dorsal surface: it is more or less apparent in Haplohymenium ( Fig. 8E View Fig ). There are probably no cases where papillae are present on the ventral surface but absent on the dorsal, although sometimes the ventral papillae may be larger than the dorsal: Anomodontella provides one such example ( Fig. 10A, B, G View 20 µm ). A quite separate case consists of papillae in the corners of the ontogenetic sectors in Herpetineuron , comparable with e.g. Atrichum , but rare in pleurocarps.

In general, variation in the pattern of papillosity in genera and families is rather gradual, and the placement of the multipapillose Heterocladium in the Lembophyllaceae and Pseudanomodon in the Neckeraceae (discussed in the present paper) looks fairly odd, and further studies are needed to understand what lies behind this. In any case, our knowledge of variation in papillae structure is still far from complete and future study will certainly reveal many interesting details for a deeper understanding of pleurocarp evolution.