Phylogeny and taxonomy of Catenularia and similar fungi with catenate conidia

Abstract The genus Catenularia (Chaetosphaeriaceae) was reviewed, and its relationships with morphologically similar fungi were evaluated using molecular and morphological data. Eleven species are accepted, four of which have been verified with molecular DNA data. The correct epithet ‘cupulifera’ is proposed for the type species C. cupuliferacomb. nov. Four other combinations are proposed, namely C. catenulatacomb. nov., C. elsikiicomb. nov., C. minorcomb. nov. and C. novae-zelandiaecomb. nov.Catenularia is an uncommon fungus inhabiting mainly decaying bark, wood and bamboo culms of various hosts and shows a widespread geographical distribution. It is circumscribed for fungi with mononematous, macronematous, simple conidiophores with terminal monophialides, usually accompanied with capitate hyphae. The conidia are aseptate, brown, cuneiform to rounded-obconic with an angular outline, adhering in chains. The diagnostic values of taxonomic characteristics of capitate hyphae and conidia (i.e. colour, shape in transverse section, setulae and formation) at the generic level were evaluated. An account of morphology, taxonomy and phylogeny of species accepted in Catenularia is provided. Based on ribosomal DNA sequences, Chalarodes obpyramidatasp. nov., characterised by catenate, angular, hyaline conidia with apical setulae, is revealed as closely related to Catenularia. The new genus Fuscocatenulagen. nov. is proposed for catenularia-like fungi having pigmented conidia with protracted maturation and round outline, with two species accepted, F. submersacomb. nov. and F. variegatacomb. nov. A new species Nawawia antennatasp. nov. is introduced and Nawawia is compared with morphologically similar taxa.


Introduction
Catenularia (Saccardo 1886) is one of the oldest genera classified in the Chaetosphaeriaceae. In April 1886, Saccardo introduced 'Catenularia Grove in litt.' with two species, 'C. simplex Grove in litt.' and C. atra (= Spadicoides atra, Hughes 1958), of which C. simplex is regarded as the type (Clements and Shear 1931). Grove (1886) intended the genus to be monotypic, and later that year published Catenularia again with C. simplex as the only species observed on wood in the United Kingdom. However, C. simplex has previously been described by Berkeley and Broome (1871) as the presumed but nameless conidial state of Sphaeria cupulifera on decaying elm roots also in the United Kingdom. The species was illustrated with pigmented conidiophores arising singly from ascomata and in tufts around them, with a funnel-shaped collarette and cuneiform, dark brown, aseptate conidia adhering in chains. The anamorph was named Psilonia cuneiformis by Richon (1877) based on a collection on wood in France and later transferred to the monotypic genus Psiloniella (Costantin 1888). Mason (1941) concluded that P. cuneiformis and C. simplex are conspecific and accepted P. cuneiformis in Catenularia with C. simplex listed as a synonym. De Seynes (1886) and Booth (1958) confirmed that S. cupulifera (= Chaetosphaeria cupulifera, Saccardo 1883) and C. cuneiformis belong to the life cycle of the same species (Fig. 1). Booth (1958) noted that the conidiophores develop from the modified outer cells of the ascomatal wall and arise from hyphae at the ascomatal bases. Linder (1933) erected Haplochalara based on H. angulospora for fungi morphologically similar to Catenularia and compared it with Chalaropsis and Thielaviopsis based on the similar pigmented, phialidic conidia in chains. Both latter genera are currently accepted in the Ceratocystidaceae ). Mason (1941) was the first to suggest the remarkable similarity of H. angulospora to Catenularia and transferred it to the latter genus. Hughes (1965) presented the first comprehensive treatise of Catenularia and accepted four species. The genus was circumscribed for lignicolous hyphomycetes with simple, pigmented conidiophores arising solitary or in tufts, with dark stromatic cells around their bases, accompanied by capitate hyphae and with monophialidic conidiogenous cells extending percurrently. The conidia adhere in chains; they are aseptate, brown, cuneiform to rounded-obconic in side view, polygonal in transverse section with a small, circular, thin-walled, pale area at each corner. Capitate hyphae, a term coined by Hughes (1949), were originally proposed for sterile hyphae scattered among conidiophores of Sporoschisma. These are erect, brown, septate hyphae that extend percurrently and terminate into a paler, swollen apex. The apical cell bears a mucilaginous hyaline cap or pale coloured droplets that may disappear with age. Capitate hyphae also occur on the ascomatal wall of the teleomorphs. Hughes (1965) did not accept the synonymy of Catenularia and Haplochalara. He considered capitate hyphae to be one of the main diagnostic features at the generic level, but which were missing in H. angulospora. Hughes (1965) excluded from Catenularia another nine species with ellipsoidal or globose, hyaline or slightly pig-mented conidia, different conidiogenous cell morphology and modes of conidiogenesis. Some of these species have been reclassified and are currently attributed to genera such as Chloridium, Exochalara, Gliomastix, Monilochaetes, Periconia, Spadicoides, Sporoschismopsis and Thielaviopsis (Mangenot 1952;Booth 1957;Hughes 1958Hughes , 1965Von Arx 1970;Holubová-Jechová and Hennebert 1972;Gams and Holubová-Jechová 1976;Schoknecht and Crane 1977;Rong and Gams 2000;Mbenoun et al. 2014;De Beer et al. 2014). Other authors did not follow such a narrow generic circumscription and several species without capitate hyphae were introduced in Catenularia, namely C. catenulata (Luo et al. 2019), C. hughesii (Sharma 1980), C. kalakadensis and C. malabrica (Subramanian and Bhat 1989), and C. variegata (Li et al. 2017). Admission of C. variegata in Catenularia introduced further heterogeneity into the genus. This species has a protracted maturation of conidia that are cuneiform or obovoid in the side view, but have round outline and lack typical corners with pore-like areas at the apex.
Species with the Catenularia morphotype have been named inconsistently as Catenularia or Chaetosphaeria. To date, 24 species and varieties have been referred to as Catenularia and six as their Chaetosphaeria counterparts (Berkeley and Broome 1871;Saccardo 1886;Linder 1933;Booth 1958;Hughes 1965;Sharma 1980;Holubová-Jechová 1982, 1983Subramanian and Bhat 1989;Réblová and Seifert 2003;Li et al. 2017;Luo et al. 2019). They have a saprobic lifestyle and occur on decaying bark, wood or bamboo culms in terrestrial, less often freshwater habitats worldwide. Pound et al. (2019) published Ch. elsikii, a fossil species similar to the Catenularia anamorph of Ch. novae-zelandiae. After the abolishment of dual nomenclature and subsequent changes to the International Code of Nomenclature for algae, fungi, and plants ( ICN;McNeill et al. 2012), Catenularia has never been formally accepted as a holomorphic genus, along with the correct taxonomic treatment of its type species.
The characteristics of conidia, conidiogenous cells, conidiophores and the mode of conidiogenesis are the main diagnostic traits that distinguish genera of the Chaetosphaeriaceae, while their teleomorphs are usually morphologically uniform. Among members of the family, Catenularia, Nawawia (Marvanová 1980) and Phialosporostilbe (Mercado Sierra and Mena Portales 1985) share a basic pattern of turbinate to obpyramidal, angular and aseptate conidia. The conidia of Catenularia are brown and without setulae, conidia of the latter genera are hyaline with several setulae at the apex, occasionally also at the base. Nawawia contains species with mononematous conidiophores, terminal monophialides elongating percurrently, and conidia aggregated in heads. In contrast, Phialosporostilbe has synnematous conidiophores associated with setae, terminal monophialides and conidia aggregated in heads, rarely in chains (Mercado Sierra and Mena Portales 1985;Sureshkumar et al. 2005). Nawawia and Phialosporostilbe are saprobes on decaying plant material, often submerged in freshwater, occasionally isolated from soil (e.g. Marvanová 1980;Mercado Sierra and Mena Portales 1985;Bhat and Kendrick 1993;Mel'nik and Hyde 2006;Wu and Zhang 2009;Goh et al. 2014). In characters of conidia, they closely resemble Chalarodes (McKenzie 1991) and Obeliospora (Nawawi and Kuthubutheen 1990), whose systematic placement remains unexplored. The genus Chalarodes includes fungi inhabiting decaying palm leaves, and is widespread in Australasia (McKenzie 1991). The conidia adhere in basipetal chains and are borne on terminal monophialides on mononematous conidiophores. The colonies of Obeliospora are composed of dark, acute setae accompanied by short, monilioid conidiophores with doliiform conidiogenous cells and conspicuous cupshaped collarettes. The genus accommodates species that thrive on submerged wood or plant litter in freshwater biotopes, occasionally they occur in terrestrial habitats, in South America and Southeast Asia (Nawawi and Kuthubutheen 1990;Kuthubutheen and Nawawi 1994;Wu and Mckenzie 2003;Cantillo-Pérez et al. 2018).
This study is based on nuc rDNA sequences combined with a comparative analysis of phenotypic data. It aims to evaluate the generic concept of Catenularia and its relationships with morphologically similar taxa. Another aim is to assess whether phenotypic characteristics such as the presence or absence of capitate hyphae and selected conidial features (i.e. colour, shape in transverse section, setulae and formation at the tip of the conidiogenous cell) are congruent with phylogenetic relationships.

Fungal strains, morphology and DNA extraction and PCR amplification
Specimens of Catenularia, Chalarodes, Nawawia and Sporoschisma were collected in various localities in temperate and tropical geographical areas in Cuba, Czech Republic, France, Belgium, Martinique, New Zealand, Slovak Republic and Thailand. Other specimens were obtained from the Canadian National Mycological Herbarium (DAOM, Ottawa, Canada), Farlow herbarium (FH, Harvard University, Cambridge, Massachusetts, USA), New Zealand Fungarium (PDD, Auckland, New Zealand), Herbarium of the National Museum (PRM, Prague, Czech Republic), and Herbarium of the Naturhistorisches Museum Wien (W, Vienna, Austria). Holotypes and specimens (as dried voucher specimens) were deposited at PDD and Herbarium of the Institute of Botany (PRA, Průhonice, Czech Republic). Fungal novelties were registered in MycoBank.
For morphological study, isolation and cultivation we follow Réblová et al. (2021a) and references cited therein. Axenic cultures were derived from freshly collected material. Strains were inoculated on potato-carrot agar (PCA) .
Protocols for the DNA extraction and PCR amplification followed Huhndorf et al. (2004), Hustad and Miller (2015) and Réblová et al. (2020). Automated sequencing was carried out by Eurofins GATC Biotech Sequencing Service (Cologne, Germany), Ottawa Research and Development Centre, Biodiversity (Mycology and Microbiology), Agriculture and Agri-Food Canada (Ottawa, Ontario, Canada) and the Roy J. Carver Biotechnology Center at the University of Illinois Urbana-Champaign (Champaign, Illinois, USA). Raw sequence data were analysed using Sequencher v.5.4.6 (Gene Codes Corp., USA, Michigan, Ann Arbor).

Alignments and phylogenetic analyses
In order to assess relationships of Catenularia with similar fungi, sequences of the internal transcribed spacer region (ITS1-5.8S-ITS2) (ITS) of the nuclear rRNA cistron and the large subunit 28S rDNA gene (28S) (ca. 1800 base pairs at the 5′-end) were analysed. Isolates, their sources and GenBank accession numbers of sequences generated in this study and those retrieved from GenBank and published in other studies (Réblová andWinka 2000, 2001;Fernández et al. 2006;Somrithipol et al. 2008;Shenoy et al. 2010;Magyar et al. 2011;Crous et al. 2012;Hashimoto et al. 2015;Hernández-Restrepo et al. 2016Liu et al. 2016;Lu et al. 2016;Ma et al. 2016;Yang et al. 2018;Lin et al. 2019;Luo et al. 2019;Vu et al. 2019;Réblová et al. 2020Réblová et al. , 2021a are listed in the Suppl. material 1: Table S1. Consensus secondary structure (2D) models for the ITS1 and ITS2 for members of the Chaetosphaeriaceae were built using the Ppfold program v.3.0 (Sukosd et al. 2012). The obtained 2D consensus models were further improved using the program Mfold (Zuker 2003) and RNAfold web server through the ViennaRNA Web Services (Gruber et al. 2015) and adjusted manually if necessary. The predicted 2D RNA structures were obtained in a dot bracket notation and were visualised and drawn using the program VARNA: Visualisation Applet for RNA (Darty et al. 2009).
Sequences were aligned manually in Bioedit v.7.1.8 (Hall et al. 1999). Consensus 2D structure models for the ITS1 and ITS2 were used to compare nucleotides at homologous positions (in helices and loops) and construct a reliable multiple sequence alignment. A predicted 2D model of the 28S of Saccharomyces cerevisiae (Gutell et al. 1993) was used to improve the alignment of this gene. The models were highly consistent in all species.
The ITS and 28S datasets, for which we assumed rate heterogeneity, were evaluated using PartitionFinder2 (Lanfear et al. 2017), implemented in the CIPRES Science Gateway v.3.3 (Miller et al. 2010), to find the best partitioning scheme for our datasets and to select best-fit models under corrected Akaike information criteria. Phylogenetic reconstructions were performed using Bayesian Inference (BI) and Maximum Likelihood (ML) analyses through the CIPRES Science Gateway v.3.3. ML analysis was conducted with RAxML-HPC v.8.2.12 (Stamatakis 2014) with a GTRCAT approximation. BI analysis was executed in a likelihood framework as implemented in Mr-Bayes v.3.2.6 (Huelsenbeck and Ronquist 2001). The phylogenetic analyses were performed as described in Réblová et al. (2021a).
The conflict-free single locus data sets were concatenated and the ITS-28S alignment (deposited in TreeBASE) was subjected to the phylogenetic analysis. Ninety nucleotides (nt) at the 5′-end of 28S were excluded from the alignment because of the incompleteness in the majority of sequences. The full dataset consisted of 2386 characters including gaps (ITS = 612 characters; 28S = 1774) and 1038 unique character sites (RAxML). For the BI analysis, GTR+I+G model was selected for both partitions. Tracylla aristata and T. eucalypti (Tracyllales) were selected as outgroup taxa.

Phylogenetic analyses
In the phylogenetic analysis of the combined ITS-28S sequences, we evaluated systematic placement of Catenularia in the Chaetosphaeriaceae and its relationships with morphologically similar taxa. The ML and BI trees were largely congruent; the ML tree is shown in Fig. 2. The Chaetosphaeriaceae included 49 well supported clades that correspond to individual genera or natural groups of species. The genus Catenularia was resolved as a monophyletic, strongly supported clade (95% ML, BS 1.0 PP) with four species, C. angulospora, C. cubensis, C. minor and C. catenulata. Catenularia resided in a statistically well supported clade at the base of the tree. This clade contained six other genera and natural groups of species, including Exserticlava vasiformis and Stanjehughesia hormiscioides, known to form capitate hyphae on ascomata of their teleomorphs. Catenularia was shown as a sister (95/1.0) to an unknown species of Chalarodes, described as Cha. obpyramidata below. Morphologically similar genera Nawawia and Phialosporostilbe were resolved as separate lineages. Chaetosphaeria submersa, superficially resembling Catenularia, was clustered in a distantly related clade containing Phaeostalagmus, and also Ch. innumera and another two Chaetosphaeria species with anamorphs with catenate conidia, i.e. Chloridium clavaeforme and Ch. phaeophorum.   Emended description. Colonies effuse, hairy to velutinous, brown, dark brown to black, mycelium partly immersed, partly superficial; composed of conidiophores, capitate hyphae and sometimes ascomata. Anamorph. Conidiophores macronematous, mononematous, solitary or in tufts, with dark stromatic hyphal cells around the bases, erect, straight or flexuous, unbranched, brown to dark brown, thick-walled, paler and thinner-walled towards the apex. Capitate hyphae scattered among the conidiophores, occasionally absent, erect, brown, extending percurrently, paler towards the apex, apical cell sterile, thin-walled, subhyaline to hyaline, slightly swollen, broadly rounded with a hyaline mucilaginous cap that may disappear with age. Conidiogenous cells integrated, terminal, monophialidic, extending percurrently, cylindrical, subcylindrical or somewhat lageniform, brown, conidia produced successively; collarettes cup-or funnel-shaped, brown, smooth or slightly roughened, margin entire or frayed. Conidia cuneiform, obclavate, rounded-obconic to broadly obovoid in side view, with an angular outline when viewed from above with 3-6 blunt corners, broadly rounded to flattened at the apex, truncate at the distinctive, hyaline basal hilum, with a small, circular, thin-walled, pore-like area visible in the cell wall at each corner, sometimes with a visible central pore at the base, aseptate, hyaline when young, fuscous, fulvous, brown to dark brown at maturity, thick-walled, smooth; formed singly, adhered in basipetal chains, occasionally in clusters. Teleomorph. Ascomata perithecial, nonstromatic, superficial, globose, subglobose to conical, papillate, glabrous occasionally with a powdery layer that disappears with age, sometimes covered with conidiophores and capitate hyphae. Ostiolar canal periphysate. Ascomatal wall carbonaceous, twolayered. Paraphyses persistent, branching, anastomosing, hyaline, longer than the asci. Asci unitunicate, short-stipitate, apical annulus non-amyloid, with eight ascospores. Ascospores fusiform, transversely septate, hyaline, smooth, without mucilaginous sheath or appendages.

Taxonomy
Habitat and geographical distribution. Saprobe on decaying bark, wood and bamboo culms of various hosts. Members of Catenularia have a worldwide distribution in temperate, subtropical and tropical geographic areas.
Notes. Hughes (1965) considered capitate hyphae to be an important diagnostic characteristic of Catenularia. These structures have long escaped attention, and mycologists began to notice them only after they were described by Hughes (1949). We studied holotype material of several species and original descriptions and illustrations to examine and trace this character in Catenularia. Capitate hyphae have not been mentioned in the original descriptions of C. cupulifera (Berkeley and Broome 1871;Richon 1877;Grove 1886). In studying collections of this species, we observed a variation in the presence of capitate hyphae. In some specimens, capitate hyphae are abundantly present, but may be scarce and difficult to find in others. Revision of the holotypes of C. cuneiformis var. minor (Holubová-Jechová 1983) and Ch. trianguloconidia (Réblová and Seifert 2003) not only revealed that both fungi are conspecific, but also led to the discovery of capitate hyphae, although they were not mentioned in the protologues of either species. They are scattered among conidiophores and easy to overlook. Phylogenetic analysis of several Catenularia representatives with capitate hyphae (C. cubensis and C. minor) and those without them (C. angulospora, C. catenulata) provided compelling evidence to consider these species congeneric.
In this study, we present a taxonomic circumscription of Catenularia using molecular and phenotypic data. The generic concept has been emended and species with and without capitate hyphae are accepted in Catenularia. We were unsuccessful in obtaining C. cupulifera into axenic culture from fresh material. The available nontype strain CBS 419.80 of this species is a contaminant (In the Blast search, ITS and  Sharma (1980) C. kalakadensis

Morgan-Jones
Sporoschismopsis simmonsii (Morgan-Jones) Hol.-Jech. & Hennebert 28S sequences derived from this strain showed 100% identity with sequences of various strains of Calycina citrina.). Eleven species are accepted in Catenularia and listed below, four of which have been verified with molecular DNA data. Other species are accepted based on morphological similarity, but have to be confirmed as members of Catenularia by molecular data. So far, the teleomorph has been observed in C. cubensis, C. cupulifera, C. minor and C. novae-zelandiae. Catenularia variegata (Li et al. 2017) is excluded from Catenularia and transferred to a new segregate genus Fuscocatenula in this study. Disposition of Catenularia and morphologically similar taxa previously attributed to the genus is presented in Table 1.

Key to
Habitat and geographical distribution. Saprobe on dead culms of Bambusa sp., decaying wood of Fagus sp. and other unknown hosts in freshwater and terrestrial habitats. It is known in China, India and the USA (Linder 1933;Sharma 1980;Luo et al. 2019 as C. cubensis).
Notes. For additional description and illustration, see Luo et al. (2019, as C. cubensis). Hughes (1965) revised the type material of H. angulospora, and despite the striking similarities to other Catenularia, he kept the species in Haplochalara due to the absence of capitate hyphae. Sharma (1980) described C. hughesii on dead bamboo culms in India with pale brown to brown conidia 6-8 × 4.5-5.8 μm and conidiophores up to 270 × 5-7 μm. Although the holotype of this species was not available for study, a detailed morphological comparison of its original description and illustration with C. angulospora suggests that they are conspecific. Luo et al. (2019) reported this species as C. cubensis (strain MFLUCC 18-1331) from China, characterised by the absence of capitate hyphae and cuneiform, greyish-brown to brown conidia 6-8 × 4-6 μm.
Notes. Our observations of the teleomorph-anamorph connection between Ch. cupulifera and C. cuneiformis agree with those of Berkeley and Broome (1871), De Seynes (1886) and Booth (1958). Although this relationship has not yet been verified experimentally, both morphs occur together in nature. Since the anamorph and teleomorph represent two different stages of the life cycle of one organism, we propose a new combination in Catenularia based on Sphaeria cupulifera with C. cuneiformis and C. simplex as synonyms.
Notes. Catenularia elsikii was isolated from the material containing clay, charcoal and wood fragments present in the cracks of a large sample of fossil wood discovered in the United Kingdom (Pound et al. 2019). Thick-walled, dark brown conidia were the only structure that has been preserved in material dated to the Miocene. In the conidial characteristics, C. elsikii is remarkably similar to C. macrospora known from Canada and New Zealand and C. novae-zelandiae known only from New Zealand. These species share dark brown, rounded-obconic conidia with (3-)4-5 corners when viewed from above. In addition, C. elsikii and C. novae-zelandiae have a visible pore at the basal hilum. Conidia of C. elsikii (23.1-24.4 μm high, 20.8-23.9 μm wide with a basal scar 3-4 μm wide) are longer and wider than those of C. novaezelandiae, but shorter than those of C. macrospora. For detailed comparison, see notes to the two latter species.
Notes. For descriptions and illustrations, refer to Subramanian and Bhat (1989) and Xia et al. (2013). Catenularia kalakadensis is unique among other species in conidia with six blunt corners when viewed from above. It resembles C. cubensis but differs in the absence of capitate hyphae and wider conidia (6-7 μm) with more corners at the apex (Subramanian and Bhat 1989). Habitat and geographical distribution. Saprobe on decaying wood, known only in New Zealand (Hughes 1965).
Notes. Catenularia longispora is well recognisable by narrowly rounded-obconic, brown to dark brown conidia that are the longest in the genus, 27-45 μm long, 16.8-24 μm wide at the apical end, 7-10 μm wide at the basal hilum, with usually three blunt corners when viewed from above (Hughes 1965).

Catenularia malabarica Subram. & Bhat,
Habitat and geographical distribution. Saprobe on decaying wood of Magnolia liliifera and an unknown host, known only in India and Thailand (Subramanian and Bhat 1989;Kodsueb et al. 2008).
Notes. For characteristics in culture, see Réblová and Seifert (2003). The apparent similarity of C. cuneiformis var. minor (Holubová-Jechová 1983) and Ch. trianguloconidia (Réblová and Seifert 2003) and their habitat on dead bamboo culms prompted a revision of both species. Examination of their holotypes revealed that they are conspecific. Additionally, we discovered capitate hyphae in the type material of both species, although they were not described in the protologues. They are scattered among the conidiophores and easy to overlook. The hyaline gelatinous cap around the swollen apex of the capitate hyphae was not observed. Conidia slightly vary in size and colour, and often smaller and pale brown conidia occur together with slightly larger and darker brown conidia.
Holubová-Jechová (1983) distinguished var. minor from var. cuneiformis (= C. cupulifera, this study) in shorter collarettes, smaller conidia and the absence of capitate hyphae. Based on their different morphology, a new combination for var. minor is proposed at the species level with Ch. trianguloconidia reduced to synonymy.
Notes. The specimen PDD 81883 of C. novae-zelandiae was isolated in axenic culture (Fig. 7O-Q). In vitro, conidia were paler than those from nature and broadly rounded-obconic. Unfortunately, the isolate is no longer viable. The other collection PDD 119362 has conidia slightly larger 17.5-21 × 18-19 μm, 5-6 μm wide at the truncate base. In both specimens, we observed several conidia with minute hyaline appendages arising from the pale, circular, thin-walled areas in the cell wall (Fig. 7K).
Habitat and geographical distribution. Saprobes on dead leaves of Freycinetia spp. (Pandanaceae) and decaying wood, known only in Australasia in New Caledonia and New Zealand (McKenzie 1991; this study).
Notes. The genus Chalarodes, typified with Cha. bisetis, was erected for dematiaceous hyphomycetes observed on leaf litter of Freycinetia spp. in New Zealand and New Caledonia (McKenzie 1991). It is characterised by mononematous, simple, dark brown conidiophores with terminal monophialidic conidiogenous cells extending percurrently and hyaline, aseptate, cuneiform, obconical to obtriangular conidia with setulae, adhered in short basipetal chains. In the protologue (McKenzie 1991), the conidia were described only in the side view with two simple setulae at the apical end. Based on the examination of newly collected material, the conidia have angular outline when viewed from above; they have (3-)4 corners with a setula inserted in each corner. Additionally, we observed sterile setae growing among the conidiophores or on the ascomatal wall. They resemble capitate hyphae of Catenularia, but the mucilaginous sheath around the apex was lacking.
To date, two species, Cha. bisetis and Cha. obconica, have been placed in Chalarodes (McKenzie 1991). A new species, Cha. obpyramidata, inhabiting decaying wood and originating from New Zealand is introduced below. The teleomorph-anamorph connection of Chalarodes is described for the first time. Based on the results of the phylogenetic study, Cha. obpyramidata is closely related to Catenularia. Réblová,sp. nov. MycoBank No: 839467 Fig. 8 Etymology. Pyramidatus (L), pyramidal, prefix ob-(L), meaning reversely, inversely, referring to the conidial shape.
Culture characteristics. On PCA: colonies 7-10 mm diam in 14d, circular, raised, margin entire, velvety-lanose, brown to dark grey-brown with whitish-grey conidial masses, reverse black. Sporulation abundant at the centre of the colony.
Habitat and geographical distribution. Saprobe on decaying wood, known only in New Zealand.
Habitat and geographical distribution. Members of the genus are saprobes on decaying plant matter in terrestrial and freshwater environments, known only in Asia in China.
Notes. Fuscocatenula is proposed as a segregate genus for fungi distantly related from Catenularia (Fig. 2), although morphologically similar. Conidia of Fuscocatenula are obovoid with a truncate base, lack an angular outline and small, circular, thinwalled pale areas in corners that are present in Catenularia. Conidia have a protracted maturation; at first they are hyaline and only later become pale brown, while still attached in a chain. Sometimes the chain consists of hyaline conidia with only one or a few mature pigmented conidia (Li et al. 2017: fig. 1;Luo et al. 2019: fig. 52). In Catenularia, conidia are also hyaline when young but mature soon and when released from the conidiogenous locus they are usually pigmented. Since Catenularia also includes species lacking capitate hyphae, this character alone is not reliable in the distinction of Fuscocatenula from Catenularia.
Two species are accepted in the genus. Li et al. (2017) introduced Catenularia variegata for a foliicolous species from China and Luo et al. (2019) described Chaetosphaeria submersa for a dematiaceous hyphomycete from submerged wood in Thailand. Both species are similar and reminiscent of Catenularia. In the phylogenetic analysis based on ITS-28S sequences, relationship of Ch. submersa and Catenularia was not supported. Molecular data of C. variegata are not available. Based on a detailed comparison of original descriptions and illustrations of both species we conclude that C. variegata is congeneric with Ch. submersa. Therefore, C. variegata is excluded from Catenularia and both species are transferred to the new genus Fuscocatenula. Habitat and geographical distribution. Saprobe on dead stems of an unidentified broadleaf tree, known only in China (Li et al. 2017).
Habitat and geographical distribution. Saprobe on decaying wood, known only in Thailand.
Notes. We were unsuccessful in obtaining N. antennata in axenic culture. The species exhibits diagnostic characteristics of Nawawia such as pigmented, mononematous conidiophores with stromatic cells around the base, terminal monophialides extending percurrently and hyaline, aseptate, obtriangular conidia with an angular outline and several simple setulae at the apex. Conidia accumulate in a slimy head. Conidiophores forming two distinct layers were also documented in N. quadrisetulata (Goh et al. 2014: figs 2, 3).
Among Nawawia species, N. antennata is well distinguished by coiled appendages and the size of conidia. Nawawia quadrisetulata is similar to the new species in conidia with mostly four angles at the apex but differs in larger conidia (30-37.5 × 22.5-32.5 μm) with longer setulae (30-57.5 μm). Nawawia antennata resembles N. filiformis (Marvanová 1980) but the latter species has conidia wider at the apex (14-18 μm) and straight appendages.

Discussion
In this study, we have reviewed the generic concept of Catenularia and its relationships with morphologically similar genera with catenate conidia using molecular and phenotypic data. The conidial characteristics, such as the colour at maturity, the outline in transverse section and presence or absence of the setulae are the main taxonomic criteria at the generic rank for distinguishing between Catenularia, Chalarodes and Fuscocatenula. Their conidia are formed successively; they are solitary and adhere in basipetal chains. These genera are compared with Nawawia, Obeliospora and Phialosporostilbe, which have similar conidia in slimy heads.
Although molecular DNA data of C. cupulifera are not available, four other morphologically similar species accepted in Catenularia were included in the analysis of ITS and 28S sequence data. Catenularia was resolved as a monophyletic strongly supported clade. Phylogenetic analysis indicates that Chaetosphaeria (Tulasne and Tulasne 1863), based on Ch. innumera with the Chloridium botryoideum anamorph (Gams and Holubová-Jechová 1976), is a phylogenetically distinct genus (Fig. 2). Therefore, Catenularia is proposed as the generic name for a morphologically well-delimited group of species whose teleomorphs were previously attributed to Chaetosphaeria. The correct epithet of the type species of Catenularia is 'cupulifera' based on Sphaeria cupulifera 1871, the earliest available epithet at the species rank; C. cuneiformis 1877 and C. simplex 1886 are reduced to synonymy. Catenularia is delimited to fungi with pigmented conidiophores arising singly or in tufts, usually accompanied by capitate hyphae, terminal monophialidic conidiogenous cells extending percurrently and flared collarettes. Conidia are pigmented, aseptate, thick-walled, formed successively from the conidiogenous locus and usually adhere in chains. They are cuneiform to roundedobconic in side view with several blunt corners when viewed from above, each with a small, thin-walled, pore-like area. The associated teleomorphs have perithecial ascomata, unitunicate 8-spored asci, persistent paraphyses and hyaline, fusiform, transversely septate ascospores. Catenularia grows on decaying bamboo culms and bark and wood of various hosts in terrestrial or freshwater habitats worldwide.
Eleven species are accepted in Catenularia, four of which have been verified with molecular DNA data. One of the accepted species, C. elsikii, is a fossil fungus. The conidia were preserved in a sample of fossil wood, dated to the Miocene, found in the United Kingdom (Pound et al. 2019). The substrate indicates a similar habitat as in the current species. Microscopic fossil fungi are difficult to identify, especially when only spores or fragments of reproductive structures are preserved (Taylor et al. 2015). Fortunately, Catenularia conidia represent a distinctive morphotype, which allows reliable identification. The majority of species of the Chaetosphaeriaceae have hyaline, thinwalled conidia and ascospores, which will likely disintegrate in the fossilized samples. On the other hand, thick-walled and heavily pigmented fungal reproductive structures are randomly present in fossil material (Pound et al. 2019). Apart from Catenularia, Adautomilanezia , Ellisembia, Stanjehughesia (Subramanian 1992), and Sporoschisma (Berkeley and Broome 1871;Hughes 1966) of the Chaetosphaeriaceae also have thick-walled and melanised conidia that may occur in fossil material or palynological preparations. Hughes (1965) suggested that conidia of Catenularia may germinate through the inconspicuous, thin-walled areas in the cell wall in corners. In the newly recorded specimens of C. novae-zelandiae, we observed several conidia with rudimentary hyaline appendages growing from these pore-like areas (Fig. 7K). This feature has not been recorded in any other Catenularia species. However, we rule out the possibility that these appendages are germinating tubes after comparing the figure in Luo et al. (2019: figure  47l) depicting germinating conidium. The presence of rudimentary conidial appendages in Catenularia may reflect its newly revealed phylogenetic relationship.
In the ITS-28S phylogeny, Chalarodes was shown as a sister to Catenularia with high statistical support. Their close relationship is also supported by similar morphologies. Chalarodes differs from Catenularia in conidia that are hyaline at maturity and have simple setulae at the apical end. Although McKenzie (1991) described conidia of two Chalarodes species from the side view only, examination of our material revealed that the conidia are turbinate to obpyramidal with an angular outline. The discovery of rudimentary setulae in C. novae-zelandiae provides a new perspective on this characteristic. Although setulae persist in Chalarodes, the appendages in Catenularia were lost during evolution or never evolved, except in the discovered case. However, the systematic placement of C. novae-zelandiae has yet to be confirmed with DNA sequence data. Our observations of Cha. obpyramidata in culture (Fig. 8O-T) correspond to those of Marvanová (1980) on Nawawia filiformis. In both species, conidia that formed in culture lack setulae.
Fuscocatenula is proposed for fungi similar to Catenularia and readily distinguished by pigmented conidia with protracted maturation, round in transverse section, lacking minute pore-like areas at the apical end, and the absence of capitate hyphae. In the phylogenetic analysis, Fuscocatenula was shown as a separate lineage, related to several Chaetosphaeria with hyaline or slightly pigmented conidia formed singly or in chains (Gams and Holubová-Jechová 1976). Its closest relatives are Ch. mangrovei with an unknown conidial state, and Ch. innumera. Chloridium botryoideum, the anamorph of Ch. innumera, forms hyaline ellipsoidal conidia arranged in imbricate chains or large heads on sympodially elongating conidiogenous cells. Phaeostalagmus cyclosporus and two Chaetosphaeria species with Chloridium anamorphs are shown as a sister subclade to Fuscocatenula. Chloridium clavaeforme and Chl. phaeophorum belong to the section Gongromeriza and resemble Fuscocatenula in slightly pigmented, short-cuneiform or dacryoid conidia forming chains or slimy droplets. Phaeostalagmus, on the other hand, represents a different phenotype. Its conidiophores are branched with lateral or terminal monophialides producing hyaline, ellipsoidal conidia in slimy heads.
Capitate hyphae (Hughes 1949) are a prominent characteristic that occurs in several members of the Chaetosphaeriaceae. They accompany conidiophores of Catenularia and Sporoschisma; they are scattered on the substrate or more frequently grow in tufts among the conidiophores or on ascomata of their teleomorphs. Capitate hyphae also occur on ascomata of Ch. capitata, the teleomorph of Exserticlava vasiformis, and Ch. conirostris (Sivanesan and Chang 1995;Fernández and Huhndorf 2005). Similar setae with a swollen apical cell but without the mucilaginous cap were observed on and around ascomata of the teleomorph of Cha. obpyramidata (this study). The presence of analogous structures have been described in the teleomorph of Stanjehughesia (Réblová 1999); they cover ascomata and their apical part, separated by a septum, is formed by an amorphous, subhyaline, clavate to almost triangular globule. All these genera, except for Sporoschisma, clustered as members of a robust clade at the base of the family tree.
Because of its mononematous conidiophores and hyaline, tetrahedral conidia with setulae arranged in corners at the apical end, Chalarodes appears similar to Nawawia (Marvanová 1980). Nawawia encompasses aero-aquatic fungi that form effuse, hairy colonies on decaying wood, bamboo culms and petioles. It is distinguished from Chalarodes by conidia that do not adhere in chains; instead they are single or accumulate in heads at the tip of the conidiogenous cells. Conidiophores often have small stromatic hyphal cells around the base. Nawawia accommodates five species of which only four, namely N. antennata, N. filiformis, N. quadrisetula, N. sasae-kurilensis, correspond to the generic concept based on N. filiformis (Marvanová 1980;Mel'nik and Hyde 2006;Goh et al. 2014; this study). The new species N. antennata resembles N. quadrisetula (Goh et al. 2014) in characters of conidiophores and conidia but differs in that the conidia are smaller and the setulae are coiled. Unfortunately, living culture or molecular data are not available to confirm its relationships. Nawawia oviformis (Peng et al. 2016) does not fit the circumscription of the genus; it has conidia with a round outline in transverse section with setulae arranged irregularly over the whole surface. These characteristics are typical of Bahusutrabeeja (Subramanian and Bhat 1977) and N. oviformis would be better placed in this genus. In the ITS-28S phylogenetic tree (Fig. 2), Nawawia and Bahusutrabeeja form separate lineages. Three species originally attributed to Nawawia have been reclassified and placed in other genera as Neonawawia malaysiana (Yang et al. 2018), Obeliospora nitida (Cantillo-Pérez et al. 2018) and Phialosporostilbe dendroidea (Yang et al. 2018). Neonawawia is particularly interesting by its formation of sporodochial conidiomata and hyaline to light brown conidiophores; it resembles Nawawia only in the characteristics of conidia. Based on phylogenetic evidence, its placement has been confirmed outside the Chaetosphaeriaceae (Yang et al. 2018).
Hyaline, turbinate conidia with an angular outline and apical setulae represent an uncommon morphotype in the Chaetosphaeriaceae. Apart from Chalarodes and Nawawia, similar conidia borne on monophialides occur only in species of Phialosporostilbe. The latter genus is distantly related to both genera and is distinguished by synnematous conidiophores associated with setae, conidial setulae occasionally formed at the base and a chloridium-like synanamorph (Mercado Sierra and Mena Portales 1985;Bhat and Kendrick 1993). The synnemata are indeterminate and although in most species the stalk is formed by compact conidiophores that climb upwards along the seta and diverge at their fertile apices, the arrangement of conidiophores of P. gregariclavata (Shirouzu and Harada 2004) is unusual within the genus. The central setiform conidiophore is accompanied by a group of shorter, parallel conidiophores that are solitary or tightly adhering to each other and may fuse. Therefore, the conidiophores of P. gragariclava may be interpreted as a poorly developed synnemata (Shirouzu and Harada 2004: fig. 10). In the characters of conidiophores, P. gregariclavata resembles members of Nawawia.
In characteristics of conidia, Chalarodes, Nawawia and Phialosporostilbe are comparable with Obeliospora, whose systematic placement remains unknown. The genus was emended by Cantillo-Pérez et al. (2018) and is readily distinguished by the absence of stromatic hyphal cells, and the presence of dark acute setae accompanied by monilioid conidiophores with terminal doliiform conidiogenous cells and flared, cup-or funnel-shaped collarettes. The conidia vary in shape ranging from round-tetrahedral, conical, pyramidal to subglobose and are hyaline, although in some species older conidia become light brown.
Although we emphasised characteristics of conidia in chains or heads to support delimitation of Catenularia, Chalarodes and Nawawia, we should look at this diagnostic trait with caution. For example, in C. minor conidia adhere in chains but in older parts of the colony conidia may form clusters. The chains break into smaller fragments, which appear as a cluster at the tip of the conidiogenous cell. In microscopic preparation, the chains readily break up into solitary conidia (Fig. 6E, L). A similar variability occurs in Phialosporostilbe. Although the majority of species have conidia arranged in slimy heads, the conidia of P. catenata form chains (Sureshkumar et al. 2005). Réblová et al. (2011) discussed this phenomenon using the example of Monilochaetes camelliae observed with an ESEM (Environmental Scanning Electron Microscope). The authors showed that there is a continuum from conidial chains to slimy heads on the phialides in culture. Chloridium is another example, e.g. Chl. clavaeforme and Chl. virescens, in which chains, cirrhi, and slimy heads can all be observed in one species in culture (Gams and Holubová-Jechová 1976;pers. obs.). It is apparently caused by the osmolarity of the medium that may affect the proportion between chains and heads.
The present investigation contributes to the knowledge of Catenularia and similar fungi with catenate conidia placed in the Chaetosphaeriaceae. Sampling of other species in the genera Catenularia, Chalarodes, Nawawia and Phialosporostilbe, which have not yet been verified by molecular data, are needed to address their systematic placement.