The Viruá National Park (VNP; Figure 1) is located in the Caracaraí district, Roraima state, northern Brazil (1°19'11"N; 61°7'17"W DMS). The climate in the region is equatorial with the rainy season intercalated by a more or less short dry season, between October and March (ICMBio 2014). This protected area presents igneous volcanic or metamorphic rocks in the hills and sandy soil of fluvial, aeolian or weathering sedimentary origin in the plains (ICMBio 2014).

Figure 1. 

Viruá National Park (location). A Roraima state in Brazil B the VNP in the central-southern region of Roraima C the actual limits of protected area (black line), the area aimed to be included during extension (green line) and the collecting points [.shp files provided by IBGE and the VNP administration]

The VNP contains in its 227,011 ha different plant formations distributed in a mosaic (ICMBio 2014) (Figures 1–2): rainforest – typical forested formation of amazon region; white-sand savannas- sandy and leached, forested to grassy, hydromorphic or non-hydromorphic plain areas; and “buritizais” – flooded areas dominated by Mauritia palms. The protected area has its western boundary at the Branco River, a line drawn a few kilometers from an abandoned fragment of the BR-174 road (known as “Estrada Perdida”) as the northern and the eastern boundaries, and by the Anauá River in the southern limit. In addition to the water discharges received from rivers mentioned above, it also receives water from the Barauana River, situated to the east and beyond the limits of the VNP, from the Iruá River, in a south-north axis, and a dense network of streams within its boundaries (ICMBio 2014).

Figure 2. 

Viruá National Park: habitats and physiognomies. A–B waterbodies with turbid (A) and translucid (B) water; C–D Areas with saturated soils during rainy season (C) and dry season (D); E–G Forested (E given by K.G. Cangani), arboreal (F) and herbaceous (G) white-sand savannas (“campinaranas”).

We investigate the herbaceous and subshrubby aquatic and palustrine angiosperms. The expeditions to collect fertile botanical samples encompassed the local dry and wet seasons, between February/2010 and January/2015. We followed Fidalgo and Bononi (1989) for both the collecting and the herborization processes and vouchers are deposited mainly at INPA and UEC herbaria. When available, we sent duplicates to UFP and/or UFRR herbaria. The acronyms are according to Index Herbariorum (Thiers 2015, continuously updated). Our inventory also included specimens previously collected in the study area and deposited at INPA, MIRR and UFRR herbaria.

We chose the collecting points non-systematically but they tended to be more concentrated in peripheral areas and along “Estrada Perdida”, due to their accessibility, and in areas where aquatic and palustrine plants are abundant (white-sand savannas of hydromorphic soils, “buritizais” and waterbodies) (Figure 1). Even though the “Estrada Perdida” is outside of the current protected area, a proposal suggests the enlargement of VNP’s borders aiming for the inclusion of areas eastwards up to the margins of the Barauana River, such that the “Estrada Perdida” would be enclosed in the VNP (ICMBio 2014).

Our identifications were based on regional floras of the Amazon and Guiana regions and specialized bibliography of each family: Core 1936, Organization for Flora Neotropica 1967- , Cook 1985, Cook and Urmi-König 1985, Kral 1988, 1992, Taylor 1989, 1991, Thomas 1992, 1997, Steyermark et al. 1995 – 2005, Araújo and Longhi-Wagner 1996, Camelbeke et al. 1997, Luceño et al. 1997, Kral and Jansen-Jacobs 1998, Wanderley et al. 2001-, Prata 2002, Gil and Bove 2004, Campbell 2005, Trevisan 2005, Simpson 2006, Thomas 2006, Rivadavia et al. 2009, Souza and Giulietti 2009, Wanderley 2011, Christenhusz 2014.

Additionally, we studied images of specimens deposited at F, K, MO, NY, P, and other herbaria with online digital images. We took into account the reliability of the identifications, with preference to types, historical collections and specimens identified by specialists; when available in the literature, we queried the original descriptions and revisions of genera or entire families. Moreover, we consulted specialists when necessary.

Rivadavia et al. (2009) described Drosera amazonica Rivadavia, A. Fleischm. & Vicent. and cited the VNP among the localities of its occurrence, and although we did not collect this species, it appears in our list. Pessoa et al. (2015) and Mota et al. (2015), respectively, held two taxonomic treatments with focus on Orchidaceae and Xyridaceae to this area, and we also included the aquatic and palustrine herbs and subshrubs cited by them in our list. When it was impossible to examine directly any specimen of the species treated, we cited the correspondent treatment.

Information on geographic distribution and authors of the species were based on TROPICOS (last access sept/2012), “Lista de Espécies da Flora do Brasil” (Forzza et al. 2015) and specialized literature for each family.

We classified the species analyzed according to morphoanatomical characters and indicated the functional ecological group that they belong (“life forms”). The categories represent a continuum from less to more specialized adaptations to the aquatic environment: palustrine plants (growing in saturated soils); emergent plants; rooted with floating leaves; free floating; and submersed plants (with life cycle entirely or partially under water) (Cronk and Fenessy 2001, Chambers et al. 2008).

For floristic comparisons, we unite some representative lists from northern Brazil and other adjacent floras, namely: savannas from northern Roraima (Miranda and Absy 1997), Várzea and Igapó areas (Piedade et al. 2010, Lopes et al. 2014; Junk and Piedade 1993, Conserva et al. 2008), Guiana Shield (Funk et al. 2007), and Northern Brazil (Moura-Júnior et al. 2015). We made an overall comparison between the lists and no richness or diversity index was estimated (Table 1).

On the identification of A&P plants occurring at VNP, it required the examination of several references, online collections and specialists. This correct identification is important as it allows the accurate list and comparison of subsampled vegetation (white-sand savannas; Vicentini 2004) and Amazon area (Hopkins 2007) with other areas and vegetation.

We still are studying some specimens of VNP, but especially when they belong to large genera, such as Eleocharis R.Br. (Cyperaceae), this process can take considerable time in the gathering of scattered and sometimes hardly accessible references.

We observed a population of cf. Eichhornia crassipes in the VNP but all individuals were sterile. The identification of sterile specimens is sometimes imprecise but floating islands of Eichhornia crassipes were observed in the Branco River (near the urban area of Caracaraí and out of the VNP) and the sterile population shown the aerenchymatous petiole characteristic of this species (Horn 2004). So we decided to list it as cf. E. crassipes (Table 1).

A peculiar case of complex identification in the VNP is the case of Bacopa egensis (Poepp.) Pennell (currently in Plantaginaceae). The identification to genus or even family was difficult, as the floras and taxonomical works consulted did not include the respective genus or contained errors in the identification keys and/or inaccurate descriptions. Not even the description provided by Souza and Giulietti (2009) encompass the morphological variation exhibited in the specimens of the VNP (K4 C3 A3). Specialists previously classified this species in the genus Hydranthelium Kunth and it does not fit in most descriptions of Scrophulariaceae s.l. or Plantaginaceae or even Bacopa.

Our data indicated new occurrences to both Roraima state and Brazil. It includes species previously registered for the Guiana Shield, an adjacent phytogeographical unit to the north (Funk et al. 2007).

Additionally, at least seven species among the new registers to Roraima are widely distributed in other Brazilian regions, e.g., Syngonanthus cuyabensis (Bong.) Giul., Hensold & L.R. Parra (Eriocaulaceae), Utricularia gibba L., U. hydrocarpa Vahl, U. nana A.St.-Hil. & Girard (Lentibulariaceae), Nymphaea amazonum Mart. & Zucc. (Nymphaeaceae), Andropogon virgatus Desv. and Axonopus fissifolius (Raddi) Kuhlm. (Poaceae). The two new records to northern Brazil, Aeschynomene scabra and Eichhornia heterosperma, are widely distributed species (Amaral 2013, Lima and Oliveira 2013) that simultaneously occur in the Guiana Shield and other Brazilian regions (Funk et al. 2007, Forzza et al. 2015). The delay of their register may be caused by the insufficient collection effort in Amazonia region.

Among A&P plants, only a few taxa are unique to a vegetation type and/or present restricted distributions (Sculthorpe 1967), and rare species generally display larger distribution areas than those of rare terrestrial species (Santamaría 2002); although Podostemaceae is an exception worth mentioning as they are commonly endemic to a single basin or even a single waterfall (Cook and Rutishauser 2007). Endemic species may occur in white-sand savannas, the predominant vegetation type in the VNP, and such species are probably spread in the numerous isolated “islands” of this vegetation type within rainforest (Anderson 1981).

There is no species endemic to the VNP and the endemic species of white-sand savannas belonging to the A&P plants that we listed here are not restrict to Brazilian territory and occur in other northern South American countries (such as Venezuela, Guyana and/or Colombia) (Table 1). Additionally, in Brazil the apparent restricted distribution in Roraima state may be in part the result of insufficient collections in white-sand savannas. We identified some specimens collected in this phytophysiognomies in Amazonas state and not yet deposited in any herbaria, which belong to such endemic species. In fact, endemic species of white-sand savannas were recently listed in other studies of the Viruá National Park (Cabral 2011, Dávila 2011, Azambuja 2012, Cangani 2012, Pessoa et al. 2015, Mota et al. 2015). Other taxa with a more restricted distribution are families, genera or species with concentrated richness or exclusive occurrence in northern South America, or only in the Guiana Shield, e.g. Rapateaceae, Abolboda Bonpl. and Utricularia benjaminiana Oliv.

We registered in the VNP all life forms usually recognized to these plants (Cronk and Fenessy 2001, Chambers et al. 2008), namely: palustrine (e.g. Xyridaceae), emergent (e.g. Montrichardia arborescens), with floating leaves (e.g. Nymphaeae spp.), free floating (e.g. Pistia stratioides) and submerged (e.g. Mayaca spp.). It is known that the number of species decreases towards strictly submerged plants (Barrett et al. 1993), as observed in our data.

Gribel et al. (unpublished data) reported the occurrence of species belonging to two genera of free floating plants in the “buritizais” in the study area (Wolffia Horkel ex Schleid. and Spirodela Schleid.), but we did not find specimens of these genera during our expeditions or in herbaria we consulted. Free-floating species depend on the nutrients dissolved in the water column (Cronk and Fenessy 2001) and the low richness at VNP may be due to the igapó characteristics attributed to the local waterbodies (low inorganic nutrient concentrations) (ICMBio 2014, Junk et al. 2011).

Concerning to the floristic comparisons, 189 species listed here were also found in Funk et al. (2007); 69 species in Moura-Júnior et al. (2015); 50 species in Miranda and Absy (1997); 13 species in Conserva et al. (2008); nine in Junk and Piedade (1993) and four in Piedade et al. (2010) (Table 1).

There are relatively few floristic lists for wetland plants in the Amazon region (Piedade et al. 2010) and in particular the northern Brazilian region remains strongly undercollected (Hopkins 2007, Schulman et al. 2007, Piedade et al. 2010). This situation makes it difficult or even impossible for us to compare our results with other works and probably turns artifacts caused by low sampling into apparent patterns (Hopkins 2007, Piedade et al. 2010).

Another study focusing on A&P plants also carried out in the VNP is yet unpublished (Paiva 2012). In it, the collecting effort was concentrated in the PPBio (Biodiversity Research Program) grid, in a more forested area within the national park, and in an area of hydromorphic open white-sand savanna near the “Estrada Perdida” road. The authors listed 19 species, some of them present in our list. The material was identified only in part and at genus level and at the time of our visit to the herbaria in Roraima, the vouchers of Paiva (2012) were not available yet. Since we could not analyze those specimens, the different taxa there listed do not appear in our list, so further comparisons could not be made.

The similarity of the aquatic and palustrine flora of the Viruá National Park with that of the Guiana Shield is evident, only 17 species of our list are absent in Funk et al. (2007) list. One must consider the geographic proximity between the two areas. Additionally, both have common limiting conditions for many plant species, mainly the nutrient-poor soils (Janzen 1974, Prance and Schubart 1978, Anderson 1981, Steyermark et al. 1995, Funk et al. 2007).

Junk et al. (2011) classified the rivers of the study area, including the sediment-rich Branco River, as black- or clear-water due to the low levels of dissolved nutrients. The igapó and várzea differ regarding their floristic composition and as to their ecological patterns (Prance 1979, Klinge 1983, Pires and Prance 1985, Piedade et al. 2010, Junk et al. 2011). Prance (1979) refers to the exclusive occurrence of some species to each vegetation type: e.g. Bombax munguba Mart. (Malvaceae), Couroupita subsessilis Pilg. (Lecythidaceae) and Hevea brasiliensis Muell. (Euphorbiaceae) to várzea areas; and Couepia paraensis (Mart. & Zucc.) Benth. and Licania apetala (E. Mey.) Fritsch (Chrysobalanaceae) and Tabebuia barbata (E. Mey.) Sandw. at igapó areas.

Two species reported in our list, namely Bacopa egensis (otherwise collected in the Solimões River, in Central America – including in rice-fields – and even in swamps near New Orleans, Louisiana; Christenhusz 2014) and Glinus radiatus (Ruiz & Pav.) Rohr. (sometimes recorded as weed) are supposedly associated with nutrient-rich habitats. These species were only collected in a lake less than 1 km from the margin of the Branco River (Ano Bom lake, which receives floodwater from the Branco River periodically), in an area with gray clay soil. Junk et al. (2011) state that the water of the Branco River, which resembles a white-water river, has low nutrient status and it is thus chemically closer to a clear-water river. The presence of these two species suggests that the Branco River may be richer in nutrients than a clear-water river.

The comparison of our list with that of an area of várzea near Manaus (Junk and Piedade 1993) illustrates some of these differences between nonarboreal Várzea and Igapó areas; only a few species are common to both areas and in the várzea vegetation Poaceae and Araceae s.l. (predominantly Lemnoideae) are the richest families of wetland species. Another study of wetland plants at the Amazonas/Solimões River interface also revealed a low number of species in common with the VNP (Conserva et al. 2008). However, both lists (Junk and Piedade 1993, Conserva et al. 2008) are based on expeditions made during the terrestrial phase (dry season), to facilitate access to the localities, influencing the results. Thus, there is an increasing richness of weedy species not necessarily adapted to wetland conditions in both inventories.

Few studies have been carried out in areas of igapó that focus on the diversity of herbaceous and subshrubby aquatic and palustrine plants (Piedade et al. 2010). Lopes et al. (2014) present a floristic analysis of A&P plant genera in six river systems with igapó characteristics in the Amazonas state. They listed 25 families and 63 genera, from which 15 families and 24 genera occur in the Viruá National Park (Lopes et al. 2014). Similarly to the pattern observed by Lopes et al. (2014), Cyperaceae is richer than Poaceae in the VNP. Those authors did not provide a species list and comparisons are not possible.

Indirectly, Piedade et al. (2010) mention four species to igapó areas: Oryza perennis Moench, Nymphaea rudgeana G. Mey., Utricularia foliosa L. and Cabomba aquatica Aubl. As regards the yellow-flowered Cabomba Aubl. with four tepals that occurs in the VNP and in the Jufari River (see Piedade et al. 2010), we identified this species as Cabomba schwartzii Rataj. Currently, specialists list this species as a synonym of C. aquatica, typically with six tepals (Forzza 2015). The name Oryza perennis is doubtful and specialists will soon submit a proposal to reject it (Robert Soreng, pers. comm.). The specimens named O. perennis in Amazonia and the Guianas are probably O. rufipogon Griff. Therefore, we also recorded all four species mentioned by Piedade et al. (2010) for the Jufari River in the VNP; all are widely distributed species.

At northern Roraima there is other non-forest vegetation with different abiotic conditions enclosed in the matrix of Amazonian rainforest, the savannas (ICMBio 2014; IBGE 2012). Again, we know little regarding A&P plants in the savannas of Roraima and must take some care when making comparisons with the few works published, such as considering the need for a revision of the identifications published in older lists. The limitations of useful taxonomic literature, faced especially before the publication of the Flora of the Venezuelan Guayana, were considerable and sometimes researchers vaguely delimited study areas.

The floristic survey published by Miranda and Absy (1997) gathers data from various previously published inventories and results of the author’s collections in the savanna region of Roraima, being the most complete list available to this area. It contains nearly 300 species: c. 30% are from wetlands and of these, 1% in strictly aquatic habitats. We herein recorded 50 of these species for the Viruá National Park, mostly species with a large geographical distribution. More recent lists of aquatic and palustrine species from Roraima state are unpublished (Neves 2007, Paiva 2012), thus only presenting a low number of more widely distributed species.

In the updated checklist of macrophytes from northern Brazil, that is based on previous lists plus recent data collected by the authors and data contained in the SpeciesLink and “Lista da Flora do Brasil” platforms, Júnior et al. (2015) listed about 540 spp. of A&P plants, among those species only 68 are in common with our list (Table 1). As their list probably gathers information from Várzea and Igapó areas with no distinction, useful discussions cannot be made at this moment. Anyway, some care must be taken as their list was based on information of the SpeciesLink platform, (according to Moura Júnior et al. 2015), where not all specimens are correctly identified. The criterion for the inclusion of the species used by those authors is not clear. Similarly, though the “Lista da Flora do Brasil” is a great source of information about species of the Brazilian flora (as cited by Moura Júnior et al. 2015), it is still incomplete and with considerable data gaps for some taxa and geographical regions, such as aquatic and palustrine species and northern Brazil.

Despite the need for more collecting effort in the inner parts of the VNP, its flora of aquatic and palustrine herbaceous and subshrubby angiosperms is clearly connected to the flora of the Guiana Shield, an adjacent phytogeographical region to the north and geologically related to the origin of white-sand savannas, the predominant physiognomy in the protected area studied here.

Only the INPA herbarium held old specimens from the studied area and although we found some bibliography concerning wetland plants of the region, publications are scattered or the most complete refer mainly the Venezuelan and/or Guiana territory. A greater collecting effort and the revision of herbarium specimens are essential to allow for a meaningful evaluation of the similarities and differences between the white-sand savannas and other savanna areas of the Amazon. Reasonably complete lists of aquatic and palustrine plants in areas influenced by igapó rivers, white-sand savannas and other savannas in the Amazon region may also uncover floristic, biogeographical, evolutionary and ecological patterns currently obscured by the inadequate collection status.

To allow for the identification of the species here listed, our group is currently producing keys (including interactive multi-access keys), descriptions with images, taxonomic comments, geographical distribution and field observations of the listed taxa. We will provide these resources shortly, in the format of an eFlora on a website about the Viruá National Park (Costa et al. in prep.).