The peatmoss (genus Sphagnum L.) flora of South America is well-known for its high species diversity and many endemic taxa (Shaw et al. 2003, 2019, Costa et al. 2011). The majority of species are recorded from the Andean regions of Bolivia, Columbia, Ecuador and Peru, and the catchment area of the Amazon basin in Brazil, Columbia, Peru and Venezuela. These areas are characterized by high-humidity and precipitation. Among those countries, Brazil has the highest number of reported peatmosses (86 species, Costa et al. 2011), representing all subgenera, namely Acutifolia, Cuspidata, Rigida, Sphagnum and Subsecunda (Shaw et al. 2016).

Subgenus Subsecunda is especially species rich in South America compared to the rest of the world. Michaelis (2019) lists and circumscribes about 110 species worldwide of subgenus Subsecunda, and South America holds the majority of Subsecunda species reported. So far, less than ten of these species are also reported from outside this continent.

More than half of South American Subsecunda species were described between 1987 and 2002 by sphagnologist Howard Alvin Crum (1922–2002); see Buck and Anderson (2003). Most of these descriptions were based on single collections made by other collectors visiting the South American continent. Crum himself points out that this is a serious weakness in the taxonomic evaluation of his described species, and he highlighted that new knowledge based on more collections and extended research are necessary to evaluate the validity of the new species. But he also wisely claimed: ‘Meanwhile, it serves a practical purpose to give names to the species, to give them visibility rather than continue them in storage’ (Crum 1987a). New knowledge of South American Subsecunda has not improved much over the two last decades, so we fully support this view.

Taxonomic classification and nomenclature of Sphagnum taxa have been, and still is a controversial matter, including problems related to formal inter- and intraspecific classifications (species, subspecies, varieties). However, many new species descriptions are based on evidence including both morphological examination and genetical analyses (Shaw et al. 2013, Hassel et al. 2018). Likewise, description and outline of new Subsecunda species are usually based on the same solid fundament with several collections from different localities and geographical areas, for example the amphi-Pacific S. inexspectatum Flatberg (Flatberg 2005, Shaw et al. 2014b) and S. miyabeanum (Shaw et al. 2014a, b). A similar approach and new knowledge are lacking for South American Sphagnum.

The Pantepui biogeographic region of the Guiana HighlandsinVenezuela,Guyana and Brazil is characterized by about 50 isolated tabletop mountains (tepuis). Tepuis are known for their concentrations of endemic species (Huber 1988, 1992). The patterns of endemism and distribution of bryophytes of the topographic tepui islands of northern South America differ from vascular plants mainly due to higher dispersal ability of bryophytes (Désamoré et al. 2010, Costa et al. 2020). Most endemic bryophytes are assumed to originate from cold-adapted Andean ancestors dispersing to the Pantepui region (Désamoré et al. 2010). Besides the endemic species, the bryophyte flora of the Pantepui region has biogeographic elements consisting of Tropical American, Pantropical, Afro-American and globally widespread species (Costa et al. 2020). Désamoré et al. (2010) points to a relationship with the West Indian Caribbean islands whereas Robinson (1986) proposed an African or even Asian phytogeographic relationship of species in the Pantepui region. A molecular study of subgenus Subsecunda revealed a well-supported lineage comprising samples from Africa, South America and Australia/New Zealand (Shaw et al. 2008) corroborating close relationships between South America and Africa.

Among the 70 Sphagnum species reported from the South American region that includes Venezuela (Am. 4 cf. Michaelis 2019), five belong to subgenus Subsecunda: Sphagnum funkiae H.Crum, S. gracilescens Müll.Hal., S. juliforme H.Crum, S. liesneri H.Crum and S. subsecundum Nees (Michaelis 2019, GBIF 2021). Likely, the number of Subsecunda species in this region is higher as very little new knowledge of South American Subsecunda has come after the work of H. A. Crum.

After collecting Sphagnum specimens from tepui in the Chimantá massif, Bolívar region of Venezuela in 1985, T. Ahti finally had a chance to present the specimens in storage to K. I. Flatberg in 2017. One of these specimens was a homogenous sample of a subgenus Subsecunda species, not belonging to S. juliforme, the only Subsecunda species previously reported from the Chimantá Massif (GBIF 2021). Furthermore, the specimen collected has turned out to be morphologically distinct and differs from all published Subsecunda species as included by Michaelis (2019). Based on morphology, we describe the specimen as a new species named Sphagnum magniporosum (see formal description and typification below). We compare it to relevant South American subgenus Subsecunda species. Due to biogeographic connections between South America and Africa pointed out by Shaw et al. (2008) we also included comparisons with selected African Subsecunda species as outlined by Eddy (1985).

Material and methods

The specimen of Sphagnum magniporosum was collected at the tepui Macizo del Chimantá, Bolívar state, Venezuela in 1985. The tepui Macizo del Chimantá is part of the Pantepui biogeographic region of the Guiana Shield in the borderland between Venezuela, Guyana and Brazil (Fig. 1). Tepuis of this region consist of > 1.5 billion years old Proterozoic sandstone (Rull et al. 2019). The climate of the region is mild, with annual average temperatures around 20°C and a precipitation of about 3600 mm (Huber 1992). The rainy season is 10 months long and the tepuis are often wet and cloudy. The southeastern sections of the Chimantá Massif are characterized and often dominated by very extensive, open peat bogs (see map in Huber 1992: 163). As noted in Ahti (1992), the Sphagnum species are present but not a major part of the vegetation. Instead, many herbaceous or subfruticose vascular plants are the main peat-forming species, such as Brocchinia (Bromeliaceae), Stegolepis (Rapateaceae), Psychotria (Rubiaceae), Baccharis (Asteraceae) and Syngonanthus (Eriocaulaceae). In addition, tall lichens, primarily reindeer lichens (e.g. Cladonia argentea, C. confusa, C. dendroides, C. densissima, C. signata, C. steyermarkii, C. subreticulata, Cladia globosa) are very abundant (Ahti 1992). Several of the peatland Cladoniaceae species are endemic to the Guianas (Ahti and Sipman 2013).

The Sphagnum plants were investigated using morphological examination, but the ploidy level was evaluated using molecular methods. The number of microsatellites per locus has been shown to be a good method for determining ploidy levels in Sphagnum (Ricca et al. 2008, Kyrkjeeide et al. 2019). DNA was extracted from dried specimens using a small part of the shoot apex. Extractions were performed using the DNeasy 96 Plant Kit (Qiagen, Oslo, Norway) following the manufacture's protocol. We genotyped 15 microsatellite markers from two plants (Table 4; for methods see Kyrkjeeide et al. 2016).

Morphological examination and terminology

The description of the new species is based on morphological characteristics found in the examined type specimen. Basic morphology of the new species has been compared and evaluated with use of keys, descriptions and illustrations of South American and African Subsecunda species (Eddy 1985, Crum 1987a, b, 1990a, b, 1992, 1993, 1994, 1995a, b, 1997, 2001, Michaelis 2019). This published knowledge constitutes the fundament for morphological comparisons and taxonomy made here.

Studies of type collections of the compared species are not included in the present study. All the compared South American species are well described, illustrated and commented on by Crum, and summarized by Michaelis (2019). The same applies for the African species included in the study (Eddy 1985).

The S. magniporosum specimen was studied by preparation of semi-permanent slides mounted in glycerol. We examined branch and stem leaves, stem and branch cross sections, outer stem and branch surfaces and cross sections of branch leaves. Stems and leaves were stained with a methyl violet solution. The morphological examinations and measurements were performed using a transmission microscope, and photomicrographs were made applying the Leica Application suite LASD ver. 2.6 for stacking. Ten stem leaves and 10 branch leaves from the middle part of branches were measured for size.

The terminology used for description and characterization of leaf pores on the hyaline cell walls of S. magniporosum: 1) Perfect pores include circular, elliptic to more irregular perforations in the hyaline cell wall membranes. Such pores are of two types, unringed pores and ringed pores surrounded by a thickened membrane ring. Perfect pores are either free-lying from the adjacent chlorophyll cells and are illustrated as median pores (sometimes named luminal pores), or they lie rather tightly adjacent to the chlorophyll cells as commissural pores in ± continuous rows. 2) Imperfect pores are unringed and only with a membrane thinning, but without perforation. Such imperfect pores look paler when stained leaves are examined in microscope. 3) Pseudopores are without perforations, closely facing the chlorophyll cells, and are delimited by semi-circular, semi-elliptic to elongate-rectangular fibril bands connected to the inner walls, and sometimes have thin membranes. Pseudopores are usually confined to the concave leaf surface.

Figure 1.

The Pantepui ecoregion of Venezuela. < https://commons.wikimedia.org/w/index.php?curid=104644538>.

The following terminology is used for the dominating pore patterns found on the convex (dorsal) surface of the hyaline cell walls of leaves among compared species: 1) 3-seriporose cells have a mixture of commissural pores and free-lying median pores. 2) 2-seriporose cells have commissural pores without median pores. 3) 1-seriporose cells have only median pores without commissural pores. Single cell end pores are often found in addition.

Delimitation of morphologically similar species

Michaelis (2019) divides the Subsecunda species in six ‘groups’. These groups are artificial classifications, and not based on phylogenetic or taxonomical criteria. The groups are primarily based on 1) microscopic morphological characteristics connected to the number of stem cortical layers, 2) number of pore rows (2 or 3) on the hyaline cell walls of the leaves and 3) shape and position of the chlorophyll cells of branch leaves in cross section.

In this respect, S. magniporosum connects best to species group 2, characterized ‘by one-layered stem cortex and more than two rows of pores on the hyaline cells’. In practical microscopy this implies that the hyaline cell walls on their convex surface of the leaves have predominantly 2-seriporose leaf cells but with a varying element of 3-seriporose cells.

This species group includes seven South American species. Sphagnum magniporosum shares the unicorticate stem of this group. But it deviates with predominance of 2-seriporose leaf cells, combined with an element of 1-seriporose cells or cell portions with large median pores.

Despite this pore discrepancy, we have found it most relevant to primarily compare S. magniporosum with the South American species aggregated in this species group of Michaelis (2019) (Table 1). In addition, we consulted the descriptions and illustrations made for species placed in Michaelis' groups 1, 3–6, but without finding species with leaf porosity similar to S. magniporosum.

All the seven South American species are endemic to tropical South America, and their known distributions are represented by herbarium collections from single localities in Brazil (five species), Colombia (one species) and Peru (one species) and none of them are recorded from Venezuela. The collections underlying the formal types are made by several collectors from 1977 to 1987 (Table 1).

No relevant new species from South America have been described since 2002 (Crum 2002). Crum's descriptions of the seven species are the basis for the treatment in Michaelis (2019) and the illustrations in Michaelis (2019) come directly from Crum's work. Three Subsecunda species from Brazil, viz. S. pluriporosum H.A. Crum (Crum 1994), S. septatoporosum H.A. Crum (Crum 1997) and S. turgescens Warnst. (Warnstorf 1895) also share the 3-seriporose cell pattern but are placed in his species group 5, with multilayered stem cortex, by Michaelis (2019). These species also differ in other morphological characters and are not further compared with S. magniporosum.

Among the five Subsecunda species from Venezuela reported by Michaelis (2019), S. subsecundum occurs in the Northern Hemisphere with a northern mainly amphi-Atlantic distribution. This species is well separated from S. magniporosum by much smaller shoots, dimorphic branch fascicles, 2-seriporose branch leaves, stem leaves much smaller than the branch leaves and without pores on their convex surface. The occurrence of S. gracilescens in Venezuela is uncertain. It is described with small, pale shoots, multistratose stem cortex, 2-seriporose leaves and triangular to trapezoid shape of the branch leaf chlorophyll cells in cross section. These two species are left out of further comparison with S. magniporosum.

Thus, we here compare S. funkiae, S. juliforme and S. liesneri with S. magniporosum. All three are described based on single herbarium collections from Venezuela (Crum 1994, 1995a). It is noticeable that none of these species are added with recent published records from Venezuela, nor are they recorded with additional collections from outside this country.

During this study we became aware that S. magniporosum showed morphological similarity to some African Subsecunda species (Eddy 1985). Eddy divide tropical/ subtropical African Subsecunda species in two subsections, Acrosphagnum (C. Müll.) A. Eddy and Subsecunda, both are classified as sections by Michaelis (2019). The six African species belonging to subsect. Acrosphagnum are excluded from comparison with S. magniporosum, as they all possess a markedly 2–3-layered stem cortex with an enlarged surface layer with usually a large external pore and have fascicles with 2–7 branches. However, two of the species (S. capense Hornsch. and S. ceylonicum Mitt. ex Warnstorf) possess leaf cell structure approaching S. magniporosum, with 3-seriporose hyaline leaf cells (Eddy 1977). Eddy's subsection Subsecunda includes S. truncatum Hornsch. (including S. bordasii Bescher.), S. rutenbergii C. Müll. and S. africanum Wel. & Duby. The morphologically variable S. truncatum is distributed from southern and eastern Africa including the East African islands (Eddy 1985). Sphagnum rutenbergii is endemic to the East African islands (Eddy 1985), while S. africanum is recorded with scattered occurrences in montane regions from Angola to Malawi (Eddy 1985). According to Karlin et al. (2014), all three species belong to the S. africanum complex of subgen. Subsecunda, and is defined as the group of African and South American species referable to ‘lineage D’ of Shaw et al. (2008; Fig. 4). Sphagnum africanum shares many similar traits with S. magniporosum, while the two other species deviate much more in morphological details. Also, S. rutenbergii and S. truncatum have more branches in the fascicles, and they are more dimorphic and less tumid than S. africanum. Hence, only the latter is compared with S. magniporosum.

Table 1.

Survey of the seven Subsecunda species known from South America circumscribed under Species group 2 in Michaelis (2019, pp. 228–230), characterized ‘with one-layered stem cortex and more than two rows of pores on the hyaline cells’.

Results

Morphological comparison with South American species

Sphagnum magniporosum differs from all the compared species belonging to species group 2 of Michaelis (2019) in various morphological characteristics, both macroscopically and in microscopical details (Table 2). All compared species except for S. subhomophyllum and maybe S. sonsonense are smaller plants, and none are described with the marked orange plant colour of S. magniporosum.

The single branches of S. magniporosum is shared by S. garysmithii and partly by S. luetzelburgii, but they differ in several other morphological characteristics.

Sphagnum magniporosum has longer branch and stem leaves than the other species except for the near equal-sized stem leaves in S. garysmithii. As in this species, the stem leaves are large compared to the branch leaves. Stem leaf shape varies among the compared species; S. garysmithii deviates with its rounded leaves with obtuse apices, and S. subhomophyllum with its more triangular stem leaves. Differences in branch leaf shapes are less obvious.

All the compared species have stem leaf fibrillosity extending into their proximal half and usually close to the base, except for S. cribriforme with few fibrils or fibril stumps confined to distal leaf end part.

Table 2.

Morphological comparison of Sphagnum magniporosum with seven Subsecunda species from South America.

Continued.

All the species compared with S. magniporosum have 3-seriporose leaf cells present on the convex surface of the branch and/or stem leaves but with comparatively much smaller median pores. S. hegewaldii can also have an element of 1-seriporose leaf cells like S. magniporosum, and in S. subhomophyllum 2-seriporose leaf cells dominate except for some 3-seriporose cells in distal leaf part. But all these species lack the 1-seriporose cells or cell portions found in S. magniporosum with pore width equal to the hyaline cells.

The pore pattern on the concave leaf surfaces varies more. All species except S. hegewaldii, S. luetzelburgii and S. subhomophyllum lack median pores in the stem leaves, and S. cribiforme is nearly without pores. Sphagnum magniporosum is the only species recorded with almost exclusively pseudopores, while the other species have only perfect pores (S. homophyllum) or a combination of perfect pores and pseudopores (S. garysmithii, S. hegewaldii, S. sonsonense, S. subhomophyllum). The special longitudinal rectangular shape of the pseudopores found in S. magniporosum, is not reported in any of the species.

The elliptic shape of the chlorophyll cells reaching both surfaces with varyingly thickened walls in S. magniporosum is mainly shared by S. cribriforme, S. hegewaldii, S. homophyllum and S. subhomophyllum. In S. luetzelburgii we find elliptic to barrel shaped chlorophyll cells, and in S. sonsonense ovate-elliptic and more broadly exposed on the concave compared to the convex surface. Sphagnum garysmithii has rectangular chlorophyll cells reaching both surfaces.

The Venezuelan species included in the comparison (Table 3) have smaller and more slender shoots than S. magniporosum, except maybe S. juliforme. None of the three species have the orange colour of S. magniporosum . Only S. funkiae and S. juliforme are reported with stem and branch leaves approaching the length of S. magniporosum. Sphagnum juliforme is like S. magniporosum described having single branches and share the one-layered stem cortex. However, S. juliforme has very short branches, and the shape of the chlorophyll cells are rectangular (not elliptic) in cross section. Sphagnum funkiae has predominantly two-layered stem cortex, branch fascicles with four spreading branches and triangular stem leaves. Sphagnum liesneri shares the elliptic shape of the branch leaf chlorophyll cells in cross section with S. magniporosum but has small shoots and branch fascicles with mostly two branches. None of the compared Venezuelan species belong to species group 2 of Michaelis (2019). Sphagnum juliforme and S. liesneri belong to species group 1, with one-layered stem cortex and 2-seriporose branch leaf cells, and S. funkiae belongs to group 4 with multi-layered stem cortex and 3-seriporose leaf cells (Michaelis 2019). But interestingly, S. funkiae, S. hegewaldii and S. liesneri have an element of 1-seriporose hyaline cell portions in their stem leaves.

Discussion

The morphological examinations and comparisons of Sphagnum magniporosum with South American and southern African Subsecunda species sharing the combinations of selected morphological characters, support the distinctness of the new species. Sphagnum magniporosum deviates significantly from the described morphological patterns of all compared species.

The most diagnostic morphological characteristics of S. magniporosum are 1) the large size of the commissural and the median pores relative to breadth of the hyaline cells on the convex surface of the branch and stem leaves, and 2) the element of 1-seriporose cells or cell portions of large median pores on convex leaf surfaces. Only the Venezuelan species S. funkiae and S. liesneri among the compared species are recorded with an element of similar occurrence of 1-seriporose cells in distal portion of their stem leaves, but the median pores in these two species are illustrated much smaller than in S. magniporum relative to cell breadth. Another deviating morphological character found in S. magniporosum is the 1–3 minute circular pores often present on lateral walls of the branch cortical cells (Fig. 2D). These pores occur in addition to the apical pore of the enlarged retort cells of the branch cortex (Fig. 2C). Such minute, lateral wall pores are not mentioned in Crum's descriptions nor by Eddy (1985). In fact, such pores seem unreported both at genus and subgenus/section levels (Russow 1894, Åberg 1937, Suzuki 1958, Savicz-Ljubitzkaja and Smirnova 1968, Nyholm 1969, Eddy 1977, Crum 1984, Daniels and Eddy 1990, Hill 2004, McQueen and Andrus 2007, Anderson et al. 2009, Hölzer 2010, Lönnell et al. 2019, Lüth 2019, Michaelis 2019). However, such pores could be easily overlooked and may be revealed in further examination of Subsecunda species. The large leaf size with stem leaves markedly longer than branch leaves, distinguish S. magniporosum from South American species, but is a shared characteristic with S. africanum (Table 3). The taxonomic relationship to other Subsecunda species is uncertain and requires further work including molecular methods.

Table 3.

Morphological comparison of three Subsecunda species recorded from Venezuela, South America and S. africanum from Africa. Information extracted from Eddy (1985), Crum (1994, 1995a, b) and Michaelis (2019).

The descriptions and morphological characterizations of S. magniporosum are based on a single collection and reflects just one particular habitat and ecological condition. Examination of more material may confirm the distinct morphology described here or reveal a broader specter of morphological variation. The comparison made with the African Subsecunda species S. africanum has similar pitfalls. Our knowledge of the phylogenetic and taxonomic significance of traits like number of stem cortex layers and leaf porosity are insufficient, but the number of stem cortical layers with thin-walled cells (predominantly one layer versus predominantly 2–3 layers) are considered of substantial taxonomic value in Northern Hemisphere taxonomic treatments of Subsecunda species. In the S. subsecundum complex identification keys are often dichotomous in this respect (McQueen and Andrus 2007). The occurrence or not of median pores on the hyaline cell walls on the convex surfaces of leaves in addition to serial commissural pores has been paid less attention in a taxonomic context. However, the documented allodiploid species Sphagnum × triseriporum (H.Suzuki) Shaw, endemic to Honshu Island in Japan, is diagnostically separated from closely allied Japanese species by occurrence of both commissural and median cell pores on the convex surface of the stem and branch leaves (Shaw et al. 2013). In that species the median pores are much smaller than in S. magniporosum relative to cell width, and the stem is multi-corticate.

Figure 2.

Microscopic pictures of different morphological characters in branch leaves of Sphagnum magniporosum sp. nov. (A) Branch, dry, (B) branch in transverse section, (C and D) branch cortex in superficial view, (C) retort cell, (D) ordinary branch cortex cells with lateral wall pores, (E) branch leaves from middle part of branches, (F–H) cell structure on convex surface of branch leaves, (F) distal end portion, (G) mid-median portion to left, mid-marginal portion to right, (H) proximal end portion, (I) cell structure on mid-median concave surface of branch leaves, (J) branch leaves in transverse section, convex side on the top. Material: TRH-B-104740.

Table 4.

Microsatellite markers for two samples of S. magniporosum show that one allele was identified for each marker, indicating a haploid ploidy level of the species. Alleles in italics had a weak genotyping signal.

The taxonomic significance of pore characteristics in South American and African Subsecunda species has so far been neglected. Evaluation of the taxonomic importance of the leaf pore characteristics relative to other morphological attributes of comparable species in a phylogenetic and evolutionary perspective remains. Median pores are common in the subarctic–arctic North American and Eurasian S. inexspectatum (Flatberg 2005) and occur even more frequently in the closely related S. orientale (Savicz-Ljubitzkaja and Smirnova 1968; Fig. 23). Occurrence of such pores can be a good character for morphological separation of closely related species. The allodiploid species S. × perfoliatum (Savicz-Ljubitzkaja and Smirnova 1968; Fig. 24) and S. ×triseriporum (Shaw et al. 2013), likely inherited their median pores from the parental S. inexspectatum and/or S. orientale.

The clades of joint South American and African Subsecunda species found by Shaw et al. (2008, Clade B and D; Fig. 4) contains both uni- and multi-layered stem cortex species, and moreover includes species with both 2-seriporose hyaline cell and 3-seriporose hyaline cells. One subclade of clade D shows close affinity between the South African species S. africanum, S. bordasii and S. truncatum and the Brazilian species S. acutirameum and S. uleanum. Clade B included in Shaw et al. (2008, Fig. 4) includes the African S. capense, S. davidii and S. pycnocladum together with species from Australia/New Zealand, and S. sonsonense from Colombia. The observed sister group relationships between the African and South American Subsecunda species suggested that recent dispersal, rather than vicariance, explains the relationship between African and South American species in clade B and D (Shaw et al. 2008; Fig. 4).

It is a striking feature that none of the South American species compared with S. magniporosum are reported with sexual organs in the relevant publications of Crum. As a result, information about sexuality and sporophytes of the various species is generally lacking in Michaelis (2019). Eddy (1985) does not report sporophytes in S. africanum and S. rutenbergii, while sporophytes are reported as rare in S. truncatum. We know that Sphagnum spores have the ability of long-distance dispersal by air currents even across continents (Szövényi et al. 2012, Sundberg 2013, Kyrkjeeide et al. 2016). High dispersal ability of bryophyte spores may prevent speciation through on-going gene flow and lead to low rates of endemism (Vanderpoorten et al. 2010). However, Meleshko et al. (2021) found that genetic similarities between Northern Hemisphere peatmosses are likely due to incomplete lineage sorting and not recent gene flow. Speciation of South American peatmosses are so far not studied, but evolution of vicariant species through niche diversification and local adaptation could be an explanation for the high degree of endemism. A combination of the vicariant and dispersal hypotheses is an obvious option.

The Venezuelan Guayana region has the highest diversity and the highest rate of endemism of Venezuelan vascular plants (Duno de Stefano et al. 2009). About 42% of the vascular plants of the 50 topographic tepui islands of northern South America are endemic species and about 60% of these are single-tepui endemics. Costa et al. (2020) suggest that the proportion of endemic species in vascular plants in the Pantepui region may be three times higher than in bryophytes. Among bryophytes about one-third of the liverwort species are single-tepui endemics whereas Sphagnum boomii (Crum 1990b) is the only single-tepui moss species reported (Désamoré et al. 2010). Higher dispersal ability of bryophytes is assumed to explain the lower proportion of endemic bryophyte species compared to vascular plants (Désamoré et al. 2010). However, there is a possibility that also S. magniporosum could be a single-tepui endemic that evolved on the isolated mountain massive of Chimantá. If so, Sphagnum is more comparable to vascular plants with possibly many single-tepui endemics.

Until recently, Sphagnum amazonicum and S. curicuriariense were known only from their type localities, but Costa (2017) and Costa et al. (2017) reported additional sites. This demonstrates the need for field work to increase knowledge of species distributions and sexual reproduction of South American Subsecunda species. This is crucial for a better understanding of the observed distributional patterns, including endemism and disjunct occurrences within South America, as well as between pantropical species of South American and African.

Microsatellites indicate that S. magniporosum is a haploid species, as only one allele was identified for each microsatellite marker. Most of the investigated species within Subsecunda are monoploids recorded with haploid gametophytes (n = 19), but the subgenus includes several documented allopolyploid species (Meleshko et al. 2018, Shaw et al. 2019).

The only flora dealing with all species from the South American continent is the worldwide monograph of Sphagnum of Michaelis (2019). Even if his taxonomic approach is conservative and does not accept recent DNA-based works on Sphagnum taxonomy (Shaw et al. 2016, Hassel et al. 2018), his identification keys and circumscriptions of in total 293 species, are of value. Michaelis (2019) lists and circumscribes about 110 species of subgenus Subsecunda worldwide, and the South American continent is the most species rich for this subgenus.

Despite the above uncertainties, the taxonomic conclusion is still that S. magniporosum is morphological different from other species. It is unlikely to constitute a variant of the currently compared species. It is a strong need of research to clarify the taxonomic and phylogenetic relationships among neotropical South American species in subgenus Subsecunda in general, based on a combination of morphological and modern molecular methods. Ideally, extensive field work should be carried out as it would provide collections of peatmosses in their specific habitats, fresh material for genomic analyses and knowledge on distributions. Such inventories should be combined with detailed observations of sexuality, sex distributions and spore production, as well as closer examination of herbarium specimens.

Taxonomy

Figure 3.

Microscopic pictures of different morphological characters in stem leaves of Sphagnum magniporosum sp. nov. (A) Herbarium shoot, dry to left and moist to right, (B) stem leaves, (C–F) cell structure on convex surface of stem leaves, (C) distal end portion, (D) distal end, median portion, (E) mid-median portion to left, mid-median marginal portion to right, (F) cell structure on proximal end, median portion to the left, marginal to the right. 3-seriporose cells with a mixture of commissural pores and free-lying median pores are seen in the center of the photo of the marginal portion, (G) cell structure on concave mid-median portion of stem leaves, (H) stem in transverse section, (I) stem cortex in superficial view. Material: TRH-B-104740.

Key to differentiate S. magniporosum from similar South American and African Subsecunda species

Data availability statement

There are no additional data for this paper.