Phialocephala s.l. is still currently polyphyletic, both by including distantly related species and by occurrence in numerous closely related clades intermixed with species named in other genera, some of which have nomenclatural priority (e.g. Cystodendron, Mollisia, Trimmatostroma). Based on ITS sequences Grünig et al. (2002) showed that Ph. fusca, Ph. humicola (= Ph. xalapensis) and Ph. virens are outside the main Phialocephala clade. ITS BLAST queries suggest the placement of Ph. fusca and Ph. humicola in the Chaetosphaeriaceae. Morphological and phylogenetic evidence also reveal that Ph. trigonospora is not congeneric with Ph. dimorphospora and most likely belongs to Verticicladiella (Grünig et al. 2009). Acephala, a genus with two described species belonging to the PAC, is congeneric with Phialocephala and differentiated only by the lack of observed sporulation in culture (Grünig and Sieber 2005), a taxonomic choice enabled by some interpretations of dual nomenclature but now unacceptable after recent changes to the Code. Although the core clade of Phialocephala s.s. clearly defined as the Phialocephala dimorphospora clade, the presence of Mollisia cinerea-like teleomorphs in this clade cast at least some doubt on its nomenclatural priority over Mollisia until this species and the genus are epitypified (an ongoing project; A. Gminder pers comm). The apparent phylogenetic affinity of the PAC and allied taxa to Vibrissea in most phylogenies is also a major taxonomic issue that requires resolution. is further evident that Phialocephala s.s. remains to be comprehensively circumscribed, awaiting reassignment of phylogenetically closely related and unrelated Phialocephala species, and the formal transfer of species currently congeneric with Phialocephala s.s. (e.g. in Acephala and Mollisia), to stable and monophyletic genus concepts.

If Ph. dimorphospora is indeed congeneric with M. cinerea, then nomenclatural decisions will have to consider the interests of users of these names and the independent taxonomic histories of these genera. A principle of the Code is priority of publication. Mollisia, described by Karsten in 1871, clearly has priority over Phialocephala (1961). Unfortunately Mollisia is not yet robustly circumscribed and the prevailing concept is broad, polyphyletic and paraphyletic (Crous et al. 2003, Grünig et al. 2009). A definitive nomenclatural conclusion requires at least a preliminary phylogenetic investigation of Mollisia. Based on Index Fungorum, 603 names have been applied to Mollisia. The number of accepted Mollisia species after excluding synonymized taxa and subspecific identifiers is 231, excluding species of other genera that are probably congeneric or that already appear to overlap phylogenetically, such as Belonopsis, Haglundia, Nimbomollisia, Pyrenopeziza and Tapesia. Only seven Mollisia species are represented in GenBank, including only one representing an ex-type strain, M. dextrinospora (better placed in Pyrenopeziza, B. Douglas unpubl).

[Phialocephala is polyphyletic but] most well-studied species belong in the P. fortinii sensu lato (s.l.)–Acephala applanata species complex. [Note that this is phylogentically distinct from P. dimorphospora, the type species of the genus]

The PAC comprises so-called dark septate root endophytes, which are ubiquitous in the Northern Hemisphere and form complex communities in the roots of conifers and ericaceous plants. Members of the PAC are apparently restricted to roots and corresponding apothecia have never been observed in nature. Other Phialocephala species are reported as foliar and branch endophytes that form apothecia on decomposing tissues.

Phialocephala and Mollisia are polyphyletic and the delineation of these two genera remains unclear; consequently, endophytes identified by ITS sequences are often arbitrarily designated as Phialocephala, Mollisia, or Acephala. While the type species of Phialocephala, P. dimorphospora, is designated by an ex-type strain and is phylogenetically well-defined, the precise identification of the type species of Mollisia, M. cinerea, is unclear and the holotype is lost, although efforts to epitypify M. cinerea are underway (A. Gminder, pers. comm.). Understanding the relationship between Phialocephala and Mollisia is crucial for defining generic boundaries within Mollisiaceae.

Earlier phylogenetic studies using rDNA sequences reported an unexpectedly close relationship between Mollisia, Phialocephala, and the aquatic genera Loramyces and Vibrissea (Wang et al. 2006a, Raja et al. 2008). Consequently, Phialocephala is sometimes considered to belong to Vibrisseaceae (Adhikari et al. 2016, Robicheau et al. 2017). In this study, the LSU, RPB1, and TOP1 phylogenies strongly support the placement of Vibrissea outside or basal to the main Mollisia lineage (i.e. Mollisiaceae), while the ITS and LNS2 gene phylogenies place Vibrissea close to the P. dimorphospora s.s. clade or the PAC, with varying support. Evidence showing the placement of Vibrissea within Mollisiaceae based on the LNS2 phylogeny is not compelling given the overall weakly supported branches and discrepancies from other genes.

Morphological characters distinguishing apothecia of Vibrissea from typical Mollisiaceae ascomata include stipes that are often several cm long and vivid yellow hymenia in some species, filiform, multi-septate ascospores often several hundred μm long that sometimes disarticulate into part-spores (Sanchez & Korf 1966), bluing reaction of the perihymenial medullary excipulum in iodine (Baral et al. 2019), and asci bearing distinct apical caps (“nasse apicale”; Bellemere 1977, Baral 1987a). Based on these morphological differences alone, it is likely that Vibrissea should be excluded from Mollisiaceae and that the discordance observed between the individual genes are a result of long-branch attraction artefacts and/or very highly conserved, less informative gene regions. Similar discrepancies between phy- logenies using protein-coding genes (RPB1) and rDNA genes (SSU, LSU) also are reported in other groups, such as Lecanoromycetes (Hofstetter et al. 2007).

Conversely, gross morphological dissimilarities suggest a more recent evolutionary history between Vibrissea and Mollisiaceae, as Vibrissea spp. share some important mollisioid characters: paraphyses with refractive vacuolar bodies, anguillospora- and phialocephala-like asexual morphs, and a textura globulosa ectal excipulum comprised of pigmented, thin-walled, round cells. Some Vibrissea spp., such as those previously placed within Apostemidium (e.g. V. flavovirens), are sessile and somewhat mollisioid. The divergent ascospore and ascus tip morphologies in Vibrissea may be autapomorphic characters resulting from adaptations to aquatic environments, similar to the divergent ascospore, ascus, and ascomatal characters observed in Loramyces and Obtectodiscus, genera that are strongly supported in Mollisiaceae. The association of Anavirga dendromorpha and its phialocephala-like synasexual morph with Vibrissea flavovirens cultures flooded with water (Hamad & Webster 1988, as Apostemidium torrenticola) is also compelling. The phialocephala-like synasexual morph has dimorphic conidia; however, conidia are often brown and may be roughened, unlike those of Phialocephala s.s. In our study, Vibrissea flavovirens CBS 121003 produced few sparsely-branched conidiophores bearing phialides similar to those described by Descals & Sutton (1976) when agar blocks containing mycelia were floated in sterile water. Additionally, the apical cap appears to be a homoplasic character also present in Lachnum aeruginosum (=Belonidium aeruginosum) and Incrucipulum ciliare, both Lachnaceae (Helotiales) species occurring on fallen Quercus leaves (Partel 2016).

Vu et al. (2019) suggested the optimal thresholds for discriminating families using ITS and LSU were 88.51 % and 96.21 %, respectively. The ITS and LSU similarities between Phialocephala dimorphospora DAOMC 87232 and Vibrissea truncorum CBS 258.91 are 87.71 % and 95.32 %, respectively, thus placing Vibrissea outside of Mollisiaceae following these criteria (Fig. 20). However, Vibrissea flavovirens CBS 121003, which produces sessile apothecia in contrast to the stiptiate apothecia of V. truncorum, shows some conflict with the other taxa in the comparison. For example, based on the threshold values of Vu et al. (2019), V. flavovirens would be considered a distinct genus from V. truncorum and within the same family as P. dimorphospora, cf. Niptera sp., and P. scopiformis, but not other Mollisiaceae species. These conflicts probably arise from the perhaps overemphasis on the LSU region but also highlight that judgement should be applied when using formulaic approaches to estimate taxonomic boundaries. Despite that caution, the threshold values provided by Vu et al. provide a good reference point and will be discussed later. Based on phylogenetic and morphological evidence, we exclude Vibrissea from Mollisiaceae, a conclusion also supported by the recent multigene phylogenetic overview of Leotiomycetes by Johnston et al. (2019). Additional taxon sampling and sequencing is needed to determine the placement of possibly related genera such as Leucovibrissea. Recently, Pocillum was synonymized under Vibrissea based on study of the type species and an ITS phylogeny (Baral et al. 2019).