Here, we test the current generic delimitation of Rostania (Collemataceae, Peltigerales, Ascomycota) utilizing molecular phylogeny and morphological investigations. Using DNA sequence data from the mitochondrial SSU rDNA and two nuclear protein-coding genes (MCM7 and β-tubulin) and utilizing parsimony, maximum likelihood and Bayesian phylogenetic methods, Rostania is shown to be non-monophyletic in the current sense. A new generic delimitation of Rostania is thus proposed, in which the genus is monophyletic, and three species (Rostania coccophylla, R. paramensis, R. quadrifida) are excluded and transferred to other genera. Rostania occultata is further non-monophyletic, and a more detailed investigation of species delimitations in Rostania s. str. is needed. The new combinations Leptogium paramense and Scytinium quadrifidum are proposed.

Classification, cyanolichens, nomenclature, systematics, taxonomy, thallus anatomy

Collemataceae is a large group of predominantly foliose lichenized fungi commonly known as the “jelly lichens” due to their gelatinous habit. This is caused by a polysaccharide matrix around the Nostoc cyanobacterial photobionts that swells and becomes extremely gelatinous when wet. Until very recently, the generic classification of the Collemataceae s. str. was very unnatural and based solely on one character, presence (Leptogium) or absence (Collema) of a cellular cortex (Degelius 1954, 1974; Jørgensen 2007). Already Degelius (1954) questioned the monophyly of Collema and Leptogium. This was also supported by molecular phylogenies (Wiklund and Wedin 2003; Miadlikowska and Lutzoni 2004; Miadlikowska et al. 2014), and somewhat surprisingly, gelatinous genera with one-septate spores that earlier were classified in Collemataceae, were shown to belong to the Pannariaceae (Wedin et al. 2009; Otálora et al. 2010; Ekman et al. 2014; Weerakoon et al. 2018) or Arctomiaceae (Otálora and Wedin 2013). Not until Otálora et al. (2013a, 2013b) investigated the family in detail was a modern classification of Collemataceae s. str. proposed. Collema and Leptogium were confirmed as highly non-monophyletic, and Otálora et al. (2013b) instead suggested accepting 10 more or less morphologically distinct monophyletic groups from their tree, as genera. In addition to Collema and Leptogium in restricted senses, six old generic names were resurrected (Blennothallia Trevis., Enchylium (Ach.) Gray, Lathagrium (Ach.) Gray, Pseudoleptogium Müll. Arg., Rostania Trevis., and Scytinium (Ach.) Gray), and two new genera were described (Callome Otálora & Wedin and Paracollema Otálora & Wedin).

Rostania, the focus of the present study, corresponds to the Occultatum-group of Collema (Degelius 1954, 1974). It is a comparatively small genus with eight currently accepted, mainly epiphytic species, characterised by very small to medium sized (ca 0.3–5 cm in diam.) subcrustaceous to subfoliose thalli with very small apothecia (ca 0.2–0.8(–1) mm in diam.) and cuboid to oblong muriform spores. All five species included in the Occultatum group by Degelius were treated in Rostania by Otálora et al. (2013b); Rostania callibotrys (Tuck.) Otálora, P.M. Jørg. & Wedin, Rostania ceranisca (Nyl.) Otálora, P.M. Jørg. & Wedin, Rostania coccophylla (Nyl.) Otálora, P.M. Jørg. & Wedin, Rostania occultata (Bagl.) Otálora, P.M. Jørg. & Wedin and Rostania multipunctata (Degel.) Otálora, P.M. Jørg. & Wedin. In addition, Rostania laevispora (Swinscow & Krog) Otálora, P.M. Jørg. & Wedin was included in the genus. Two further species were later added, Rostania paramensis (P.M. Jørg. & Palice) P.M. Jørg. & Palice (Jørgensen and Palice 2015) and Rostania quadrifida (D.F. Stone & McCune) McCune (McCune et al. 2014). Degelius (1954) divided Collema occultatum into two varieties: var. populinum which was characterised by a squamulose, somewhat lobate thallus, and which almost exclusively grew on the bark of Populus, and var. occultatum with a granulose thallus and which occurred on many deciduous trees, including Populus. Perlmutter and Rivas Plata (2018) combined var. populinum in Rostania, as R. occultata var. populina (Th. Fr.) Perlmutter & Rivas Plata.

Otálora et al. (2013a, 2013b) included only three species (R. ceranisca, R. multipunctata and R. occultata) in their phylogenies, and thus the taxonomical position of most species has not been tested by molecular methods. As there is a substantial variation in shape and size of the lobes, apothecia and ascospores, as well as the hyphal arrangement in the thallus among the Rostania species, and as several former Collemataceae taxa have been shown to belong outside the family, the delimitation of the whole genus needs investigation. Here, we will test the generic delimitation of Rostania and investigate the relationships of any species falling outside Rostania s. str. Finally, we will note and comment on any indication of species non-monophyly, in this genus.

We produced 61 new sequences (Table 1) for the phylogenetic analyses (24 mtSSU, 15 MCM7, 22 b-tub) including 57 taxa and 1947 nucleotide positions (735 for mtSSU and 582 for MCM7 and 630 for b-tub) for the final matrix. The alignment contained 618 parsimony-informative characters (177 for mtSSU, 237 for MCM7 and 204 for b-tub). The most likely tree from the RAxML analysis is presented in Figure 1 with likelihood BS, Bayesian PP and parsimony JK support superimposed.

Figure 1. 

The most likely tree from the combined RAxML analysis based on 1947 aligned characters of mtSSU rDNA, MCM7 and b-tub from 57 specimens. Support values (Likelihood BS/Bayesian PP/parsimony JK) given when BS ≥ 70%, PP ≥ 0.90 and parsimony JK ≥ 70%. Branches receiving BS ≥ 75 %, PP ≥ 0.95 and JK ≥ 75% are indicated with a black dot. The different colour indicate different genera: blue = Leptogium, pink = Rostania, orange = Scytinium.

The analyses resulted in a topology (Fig. 1) very similar to the results of Otálora et al. (2013a, 2013b). Some of the backbone topology, however, has unfortunately no or low support. In Otálora et al. (2013b)Callome was the sister to Rostania, but in our study this relationship is not formed. All Rostania species are nested within Collemataceae, but Rostania in the sense of Otálora et al. (2013b) is non-monophyletic. Three species form a core group, which we here treat as Rostania s. str. Rostania s. str. is well supported and includes Rostania occultata (Fig. 2A), R. ceranisca, and R. multipunctata. We can conclude that R. occultata as currently circumscribed is non-monophyletic. Rostania multipunctata (Fig. 2B) shares the cuboid shape and size of the spores with R. occultata s. lat. (Fig. 3A), but the thallus differs in size (the lobes are generally larger, up to ca 2.5 cm long in R. multipunctata, while in R. occultata s. lat. they are up to ca 3 mm long). It has also accessory lobules developing from the wrinkles (Fig. 2B), which do not occur in R. occultata s. lat. The delimitation of the two varieties of R. occultata is unclear, as is the separation from R. multipunctata. Our study is not designed to study species-delimitations and we will extend our investigation of this species complex in a larger study currently in preparation.

Figure 2. 

Thallus habitus: ARostania occultata var. populina (Odelvik 1269, S), thallus lobes (arrow) B Rostania multipunctata (Poelt 8852, GZU), accessory lobules (arrow) C Rostania ceranisca (MW_HOCH020, S), accessory finger-like lobules (arrow) D Rostania laevispora (isotype of Collema laevisporum Swinscow & Krog, Tanzania, 1986, Swinscow & Krog T 3/6, O-00298), apothecium in initial stage (arrow) E Rostania callibotrys (Moberg 4431a, UPS), apothecium in initial stage (arrow). tL = thallus lobes, al = accessory lobules, Ai = apothecium in initial stage covering the top of the accessory lobules. Scale bar: 1 cm.

Rostania ceranisca, the only terricolous Rostania, is sister to the group consisting of R. multipunctata and R. occultata s. lat. In addition to its terricolous ecology, it is easily recognized by the erect accessory finger-like lobules (Fig. 2C), which grow from the edge of the main lobes. The spores in R. ceranisca differ in shape from the cuboid spores in R. multipunctata and R. occultata s. lat. (Fig. 3A) in being oblong (Fig. 3B). Although Degelius (1954) noted only four spores in the ascus, we have usually observed eight spores, even if four of them may be aborted or are at least not clearly visible when mature (Fig. 3B).

Figure 3. 

Ascospores: ARostania occultata var. populina (Llop 56060303, S), cubic spores B Rostania ceranisca (Westberg L271_PL433, UPS), oblong spores; ascus (red line) with only four mature spores visible but remnants of four aborted spores can be seen (arrows) C Rostania callibotrys (Sipman 2049, GZU), oblong spores. Scale bar: 10 µm.

Rostania callibotrys does not group with Rostania s. str. (Fig. 1), but forms an unsupported group with Enchylium. Rostania callibotrys has a comparatively distinct thalline apothecium margin, similar to some species of Enchylium. However, this is a widespread feature in the family including some species of Rostania s. str. The thallus with characteristic accessory lobules in R. multipunctata (Fig. 2B) and R. laevispora (Fig. 2D) is very similar to R. callibotrys (Fig. 2E). Rostania callibotrys also has spores that are very similar to the typical cuboid to oblong Rostania-spores in R. multipunctata and R. occultata s. lat. (Fig. 3A, B), but the spores in R. callibotrys have fewer cells (Fig. 3C) than in these species. Rostania laevispora (Fig. 2D), a rarely collected species that we did not manage to get sequences from, is very similar and likely very closely related to R. callibotrys (Fig. 2E). As there is no support for excluding these species, and no distinct morphological evidence suggests any other relationship, we tentatively leave both R. callibotrys and R. laevispora in Rostania.

We did not manage to get molecular data from R. coccophylla (Fig. 4A), a tropical and rarely collected species where the available material was too old. Although R. coccophylla is similar to R. callibotrys and R. multipunctata, the apothecia in R. coccophylla are very different in that they are convex and stipitate when mature (compared to concave and initially immersed and later sessile, in Rostania) and considerably larger compared to other Rostania species. The apothecia of R. coccophylla are similar to several species in Collema sensu Otálora (2013b), where this species originally was placed. Although we preferably would want molecular data to test the correct placement of this species, we suggest that it is re-instated in Collema, where the name Collema coccophyllum Nyl. is available.

Figure 4. 

Thallus habitus: A Rostania coccophylla (isotype of Collema coccophyllum Nyl., India, 1858, Perrotet s.n., H-NYL 42355, H–9201376) B Rostania paramensis (Palice 2796, PRA-00013999; holotype of Collema paramense PM Jørg. & Palice) C Scytinium sp. Palice 2274a D Rostania paramensis Palice 2274b. tL = thalline lobes, is = isidia. Scale bar: 1 cm.

Rostania quadrifida and R. paramensis are not closely related to Rostania s. str. Rostania quadrifida was described by Stone and McCune (2010) as Collema quadrifidum, and was later included in Rostania based on spore shape and thallus morphology (McCune et al. 2014). It differs from Rostania s. str. by having spores with fewer septa (Fig. 5A). Here it forms the sister group to Scytinium (Fig. 1), within a well-supported group consisting of Blennothallia, Lathagrium and Scytinium. Rostania quadrifida has a thallus composed by densely interwoven hyphae, and with a pseudocortex (Fig. 6A), features that do not occur in Rostania s. str., but in some species of Scytinium (similar to e.g. Scytinium intermedium and S. magnussonii; Jørgensen 1994). These similarities support including it in Scytinium, which we do below.

Figure 5. 

Ascospores: A Rostania quadrifida (McCune 2744, UPS), cubic spores with 2–5 cells B Rostania paramensis (Palice 2796, holotype of Collema paramense), ellipsoid spores with acute ends C Leptogium azureum (Tehler 3140, S), ellipsoid spores with acute ends D Scytinium sp. (Palice 2273), oblong spores, E Scytinium sp. (Palice 2274), oblong spores. Scale bar = 10 µm

The generic position of R. paramensis has been complicated to assess. Jørgensen and Palice (2012) described it as Collema paramense, based on the holotype (Palice 2796) and another sample from a second locality in Ecuador (Palice 2273). As the thallus has a pseudocortex, Otálora et al. (2013b) transferred it to Scytinium. Jørgensen and Palice (2015) later studied another sample from the second locality (Palice 2274). They concluded that the spores in the holotype must have been unusually developed, and transferred it to Rostania based on the oblong spores (similar to R. ceranisca) found in Palice 2274. Our re-examination of these three specimens, including the holotype, shows that Palice 2273 and Palice 2274 contain two distinct Collemataceae species (Fig. 4C, D). One of these (Fig. 4D), present in small amounts only in both samples, is identical with holotype of Collema paramense and is characterised by a matt dark olive thallus with a pseudocortex (Fig. 6B), and hyaline, muriform, ellipsoid spores with acute ends (Fig. 5B). This is very different from the spores in Rostania, but typical for species in Leptogium s. str. (Fig. 5C). We sequenced the holotype, and we can conclude that among the Leptogium species we have sampled, it forms a group with Leptogium azureum (the conserved type of Leptogium; Jørgensen et al. 2013) and L. denticulatum (Fig. 1). It has a thallus which is appressed to the substrate and composed by relatively small lobes (Fig. 4B) which is rare in other Leptogium s. str., and in section it has straight and unbranched hyphae which are perpendicular to the surface (Degelius 1954; Fig. 6B). This character is present in several groups in Collemataceae. It was observed by Degelius (1954) in some Collema species, and has also been noted in the newly described Leptogium antarcticum by Kitaura et al. (2018) who used the term “columnar hyphae” for the same hyphal arrangement. We have observed this hyphal arrangement in Leptogium azureum (Fig. 6C) and L. denticulatum too, but it is apparently not present in Rostania. The second species present in Palice 2273 and Palice 2274, apparently confused Jørgensen and Palice (2015) as their observation of oblong spores (Fig. 5D, E) refer to this species, which has a shiny brown thallus (Fig. 4C) and not a matt dark olive thallus as in “Rostania” paramense (Fig. 4B). The second species differs from Rostania by having a proper eucortex (Fig. 6D), and by producing isidia along the apothecium margin (Fig. 4C). The thallus is paraplectenchymateous throughout (Fig. 6D). This hyphal arrangement is present in several groups in Collemataceae, including Rostania occultata s. lat. (Fig. 6E). Already Degelius (1954) noted this hyphal arrangement in his Occultatum-group and Otálora et al. (2013b) observed the same in Blennothallia, Pseudoleptogium and in Scytinium. We sequenced also this species and we can confirm that both samples belong in Scytinium, but the species remains to be identified.

Figure 6. 

Thalli, transversal cross-sections: A Thallus with pseudocortex and densely interwoven hyphae (Rostania quadrifida, McCune 2744, UPS) B Thallus with pseudocortex and straight and unbranched hyphae which are perpendicular to the surface (Rostania paramensis, Palice 2796, holotype of Collema paramense) C Thallus with eucortex and straight and unbranched hyphae which are perpendicular to the surface (Leptogium azureum, Tehler 3140, S) D Thallus with eucortex and paraplectenchymateous throughout (Scytinium sp. Palice 2273) E Thallus paraplectenchymateous throughout (Rostania occultata var. populina, Llop 56060303, S) A–E in lactic blue C in water. euC = eucortex, PsC = pseudocortex, Hp = hyphae, pPlect = paraplechtenchyma. Scale bar = 10 µm

Here we have tested the current generic concept of Rostania and conclude that at least three of the species should be excluded and that the position of R. callibotrys and R. laevispora in Rostania is uncertain. Rostania is characterized by crustose to subfoliose thallus with initially immersed apothecia (Fig. 2D, E), which only later become sessile. The disc is concave when young and plane when older, but never convex. The spores are muriform with at least 5 cells, cuboid to oblong, but never fusiform to ellipsoid (Fig. 3A–C). Most species are comparatively small, and all lack cortex, rhizines and isidia.

Rostania includes six taxa: R. callibotrys, R. ceranisca, R. laevispora, R. multipunctata, R. occultata var. occultata, and R. occultata var. populina. Rostania occultata s. lat. is non-monophyletic and this species complex will be investigated in the near future.

This project was funded by the Swedish Taxonomy Initiative (Svenska artprojektet) administered by the Swedish Species Information center (ArtDatabanken), grant 2016-207 4.3. We would like to thank Ulrika Nordin and Fredrik Jonsson for their kind help during the fieldwork. We are grateful to the staff of the herbaria H, O, PRA and UPS for loans and the staff at S for administering our loans. Finally, we thank the staff at the Department of Bioinformatics and Genetics at the Swedish Museum of Natural History, in particular Bodil Cronholm and Wendy Solis for their skilful lab assistance; and Johan Nylander for his helpful discussions regarding selection of best-fit partitioning schemes for analyses.