Preliminary Angiosperm Checklist in an Area South of the Madeira River, Manicoré, Amazonas, Brazil

Due to the large extent of the Amazon rainforest, research has historically focused on easily accessible locations. Thus, much of this region has little information available about its richness and plant distribution. Located south of the Madeira river, Municipality of Manicoré has a high number of phytophysiognomies, which may indicate the existence of a greater diversity of plant species. Therefore, from the compilation of previously collected records and based on botanical expeditions, this study evaluated the diversity and richness of angiosperms in Manicoré. We found 801 species, 409 genera and 106 families. Our data record 47 new occurrences for Amazonas State. Of these new occurrences, 12 are also the first record for the northern region. In addition, we have identified a new vine species of the genus Mandevilla Lindl. Given the well-known sample deficiency of the Amazon region, and considering the countless anthropogenic pressures that cities south of the Madeira river have been facing, knowledge of flora becomes increasingly urgent.


Introduction
The Amazon rainforest is an extensive cluster of landscapes and ecosystems that form an exuberant mosaic of vegetation and hydrography. Its area covers 8 million km², distributed in nine South American countries (Araújo, 2008;Porto-Gonçalves, 2015). It is the largest tropical rainforest in the world, being considered the largest reservoir of biodiversity on the planet (Porto- Gonçalves, 2015), besides being the main source of all biodiversity in the neotropical region (Antonelli et al., 2018).
Despite containing all this biodiversity, much of the Amazon rainforest is still unknown and for many reasons the most of its area remain undersampled (Hopkins, 2007;ter Steege et al., 2016). Approximately 33,300 angiosperm species are listed for Brazil, of which more than 12,000 occur in the Amazon and more than 8,500 in Amazonas State (Flora do Brasil 2020, under construction). However, estimates of the total number of species are much higher (BFG, 2015;ter Steege et al., 2016;Domingos et al., 2017). Due to the large extent of the Amazon rainforest, research The botanical material was collected according to guidelines for herbarium collection (Peixoto & Maia, 2013).
To classify species habit, the definitions presented by Gonçalves and Lorenzi (2007) were considered. For the samples obtained from the herbarium database, the habits described by the collectors were considered.
To identify the species, we used specialized bibliographies and comparisons between images available in the digital herbarium of the Botanical Garden of Rio de Janeiro -JABOT (2018), CRIA (2018), and Flora do Brasil 2020 (under construction), in addition to consultations with experts. After identification, the material was incorporated into the INPA herbarium, with duplicates in the HUEFS, RB, and UPF herbariums (herbarium acronyms according to Thiers, continuously updated). Flora do Brasil 2020 (under construction) was consulted to verify species nomenclature and confirm the records for Amazonas State and by phytogeographic domain. When this was unavailable, the synonyms and spellings of the taxa were updated by consulting The Plant List (2010) database. Specimens unidentified at the species level were not included in the checklist, which contains only one voucher per species occurring in Manicoré.
The map of the location of the studied municipality ( Figure  1) was generated in ArcGIS 10.3. The map of the distribution of angiosperm collections in the municipality was generated in Quantum GIS 1.7. For the elaboration of this map, geographic coordinates available in the INPA and CRIA databases (2018) were used. However, 313 species were not included because records did not have geographic coordinates or because species were incorrectly georeferenced. The map of angiosperm records for cities south of the Madeira river was also generated in Quantum GIS 1.7, using data obtained from the CRIA digital archive (2018), which were manually analyzed and filtered.

Results and Discussion
A total of 106 families, 409 genera, and 801 species were listed for the municipality of Manicoré (Table 1). The richest family was Fabaceae (117 spp.), followed by Euphorbiaceae For Brazil (2,756 spp.), as well as for the Amazon biome (1,119 spp.) and Amazonas State (825 spp), Fabaceae is the family with the highest species richness (BFG, 2015; Fabaceae in Flora do Brasil 2020, under construction). In the Ducke Reserve Flora Project, which is the best studied area of the Brazilian Amazon, Fabaceae (still considered in the circumscription of subfamilies Mimosoideae -68 spp., Papilionoideae -66 spp., and Caesalpinioideae -54 spp.) is also the best represented botanical family (Ribeiro, Nelson, Silva, Martins & Hopkins, 1994;Hopkins, 2005). In addition to this family, all the other richest families in our study (except Malvaceae) are in the ranking of the thirty richest families in number of taxa, being also found in the Ducke Reserve (Ribeiro et al., 1994;Hopkins, 2005).
Other studies also conducted in the Amazon showed that Fabaceae has been consistently cited as one of the families with the largest number of species (Amaral, Matos & Lima, 2000;Oliveira et al., 2008;Silva, Matos & Ferreira, 2008;Pinheiro et al., 2010). The presence of representatives of Fabaceae, Euphorbiaceae, Moraceae, and Malvaceae among the richest families is a common point among the floristic surveys conducted in the Amazon Forest (Gonçalves & Santos, 2008;Sardinha, Freitas, Santos, Cruz-Junior & Santos, 2017).
A C B D Also corroborating our findings, a study of a forest fragment in southwestern Amazonia mentions Inga, Miconia, and Byrsonima as the richest genera (Oliveira, Nagy, Barros, Martins & Murta-Junior, 2015).
Of the remaining genera, 256 are represented by only one species, 68 by two species, and 30 by three species, which together amount to 44.2% of the municipality richness. The large number of genera represented by few species may reflect the high number of plant typologies (44) found in Manicoré, as already mentioned by Silva & Pereira (2005). Notwithstanding, other factors may also be considered, such as the absence of dispersers, nutrient-poor soils, among other factors, indicating that further studies should be carried out.
In Manicoré, there is a ratio of four woody species (tree, shrub, liana, and subshrub) to one herbaceous species (herb, herbaceous vine, and palms). The predominance of trees over other types of habits has already been observed in another study conducted in the Amazon Forest (Garcia, Silva, Zonetti & Romagnolo, 2011) and follows the general pattern recorded for the Amazon phytogeographic domain (BFG, 2015). In turn, the herbaceous habit is better represented in open areas, as already reported by Mota et al. (2018).
Regarding distribution, 49.2% (394 spp.) of the registered species are exclusive to the Amazon domain and 48.4% (388 spp.) occur in the Amazon and other Brazilian domains, especially Cerrado (10.2% -82 spp.), Atlantic Forest (5% -40 spp.), Pantanal (0.5% -4 spp.), and Caatinga (0.4% -3 spp.). Moreover, 29% (232 spp.) of species share between three to five domains and 3.4% (27 spp.) are cited for six Brazilian biomes. It is noteworthy that 1.5% (12 spp.) of the total species had no association with the Amazon phytogeographic domain so far (BFG, 2015). Therefore, Asystasia gangetica (L.) T. Species dispersal and exchange among different phytogeographic domains are related to numerous evolutionary processes and historical geological events (Fiaschi & Pirani, 2009;Batalha-Filho & Miyaki, 2014). Thus, the occurrence of species with disjoint distribution patterns between the Amazon and the Atlantic Forest shows a possible connection between the floras of these regions through the Cerrado in the past (Fiaschi & Pirani, 2009). In this sense, gallery forests were responsible for the connection between the two largest neotropical rainforests: the Amazon Rainforest and the Atlantic Forest (Méio et al., 2013).
Regarding origin, 96% (769 spp.) of the species are native and 4% (32 spp.) exotic. During floristic-phytosociological surveys or taxonomic reviews, it is common to find exotic plants in the study areas (Moro et al., 2012). In this regard, Althernanthera tenella Colla (Amaranthaceae), Jatropha gossypifolia L. (Euphorbiaceae), Lantana camara L. (Verbenaceae), Lippia alba (Mill.) N.E.Br. ex P. Wilson (Verbenaceae), Merremia umbellata L.Hallier f. (Convolvulaceae), Mimosa invisa Mart. ex Colla (Fabaceae), and Ricinus communis L. (Euphorbiaceae) are found in disturbed areas of Manicoré. These plants are typical of secondary succession, are able to grow in adverse conditions, and are an integral part of the urbanized landscape (Souza, Machado-Filho & Andrade, 2012). Exotic plants are more likely to be found in these areas, with only a small fraction of them being naturalized, such as L. camara and R. communis for example. Thus, the presence of naturalized species in the study area is a strong evidence of the anthropogenic influence on the environment (Schneider, 2007).
The degree of aggressiveness that an exotic species can present to the natural environment is not always known (Schneider, 2007). Considering that botanical surveys are the basis for establishing criteria for the prevention and control of possible damage to the natural environment, it is recommended that all naturalized or invasive exotic species be clearly named as such and recorded for the study area (Schneider, 2007;Moro et al., 2012).
In this checklist, as well as in Flora do Brasil 2020 (under construction), Asystasia gangetica (L.) T. Anderson (Acanthaceae), Cenchrus purpureus (Schumach.) Morrone (Poaceae), Citrus x aurantium L. (Rutaceae), and Gossypium barbadense L. (Malvaceae) are labeled as naturalized. In Cuba, these species are considered invasive, with high capacity for growth, proliferation, and dispersal, and are often able to compete aggressively for dominance of the environment. In this sense, C. purpureus and G. barbadense are examples of species that still behave as transformers of natural and agricultural ecosystems in Cuba (Prieto & González-Oliva, 2015).
For floristic and phytosociological studies, it is only interesting to report the occurrence of exotic species merely cultivated to the site if they are clearly labeled in the study description (Moro et al., 2012) During field expeditions for this work, material from Hibiscus sabdariffa L. was collected from a disturbed area of secondary forest (Figure 3a-b). Coelho and Amorim (2019) found that this fact corresponds to an indication of naturalization of this species in the Brazilian Amazon.
Prior to the study by Coelho and Amorim (2019), Hibiscus sabdariffa was only recorded as cultivated (Esteves, Duarte & Takeuchi, 2014), being absent from the records of Brazilian angiosperms (BFG, 2015;Hibiscus in Flora do Brasil 2020, under construction). In this checklist, as in the study by Coelho and Amorim (2019), this species is labeled as naturalized and recorded as new occurrence for Brazil (Table 1). In addition to the new occurrences of exotic species, 39 new occurrences of native species were recorded for Amazonas State, eight of which are also new occurrences for the northern region (Table 1). It is noteworthy that of the new occurrences for Amazonas State, 20.5% (8 spp.) do not occur in the Amazon phytogeographic domain. However, they used to occur exclusively in the Cerrado or associated with this domain, showing great floristic heterogeneity for the study area. In the southern area of the Amazon, fields or scrublands are supposedly growing due to a shift in a 200-kilometer-long segment of the Madeira river to the east. This shift occurred a few years ago from a rearrangement of tectonic faults, changing the location of many of the right bank tributaries (Pivetta, 2011). A new type of vegetation emerges over the old beds of these rivers that were buried with sandy sediments, forming fields and scrublands in the Amazon (Pivetta, 2011).
Our data indicate that the first botanical collections in Manicoré were performed in the early twentieth century, precisely in 1923 (n = 1 collection). By 1970, only 32 species had been collected. In this sense, collection expeditions were amplified from the 70's (n = 47) and 80's (n = 152), decreased in the 90's (n = 1), and expanded again in 2007 (n = 209) and 2018 (n = 176). Although the 1980s accounted for the largest collection peak of the last century, no study addressed the floristic composition of the municipality. However, these collections constituted works of greater geographical scope, such as the RADAMBRASIL Project.
Despite collection efforts made in the last century, about 70% of the species presented in this study were collected in the 21 st century, with two major collection peaks in 2007 and 2018. Botanical collections performed in 2007 are part of an ethnobotanical survey (Junqueira, 2008). Collection peaks between 2018 and 2019 are the result of 28 botanical expeditions made for this checklist, which accounted for 25.2% of the total species listed for Manicoré, making the sites of these collections become the areas with the highest record of angiosperm species for the municipality (Figure 4).
Collection gaps, mainly in the central region of Manicoré, indicate the lack of research in much of the municipality. For botanical collections, forest areas with easier access are generally better researched (Hopkins, 2007). Collection records for Manicoré, are concentrated near the urban area, on the banks of the Madeira river and on part of the BR 319 Highway, where the main communities of the municipality are located (Figure 4).
In the expeditions carried out in this study, we collected 240 individuals from 63 families, 145 genera, and 202 species (Figure 4). In just 90 h of field sampling, besides finding evidence to prove the naturalization of Hibiscus sabdariffa (Coelho & Amorim, 2019), it was also possible to identify a new species of Mandevilla Lindl. (Apocynaceae) endemic to the southern Amazon region (Coelho et al., in press) (Figure 3c). Therefore, we estimate that the areas that have collection gaps in Manicoré also have the potential to harbor new taxa for science.
During floristic surveys, it is common to discover new occurrences (Ivanauskas, Monteiro & Rodrigues, 2004;Lopes, Ribeiro, Rodrigues, Cabral & Silva, 2014) and occasionally new taxa for science (Baitello, Arzolla & Vilela, 2017). This shows part of the advances made in recent years, but mainly indicates how much the Brazilian flora still needs to be known (Peixoto & Morim, 2013). The new species of Mandevilla was found on the edge of a newly opened road in a primary forest fragment located in the municipality countryside ( Figure 3d). This new species is described as Mandevilla manicorensis C.A. Coelho, B.S. Amorim & J.F. Morales (Coelho et al., in press.). The species shows foliate bracts and hypocrateriform corolla and is part of the Exothostemon group. In this group, twelve species have hypocrateriform corolla and only four species have foliate bracts. Easier road access contributed to this botanical discovery. However, ease of access may compromise local flora in the future, with the expansion of the municipality urban center (Salles, Grigio & Silva, 2013).
Among the cities south of the Madeira river, Beruri, Tapauá, Apuí, Novo Aripuanã, and Manicoré are, respectively, those with the largest gaps in angiosperm collections (Coelho & Amorim, unpublished data), which could be indicative of low richness ( Figure 5). Nonetheless, collection gaps in these cities are due to low sampling effort rather than absence of species, as we can find a large number of species in neighboring areas of similar forests. Thus, we can affirm that these areas are subsampled and consequently prone to contain species not yet identified or described (Hopkins, 2007;ter Steege et al., 2016).
Comparing Manicoré to neighboring municipalities, some factors help to understand why Humaitá is less subsampled. The municipality of Humaitá has access roads through Amazonas State and Rondônia. In addition, it houses a larger number of higher education institutions, which facilitates research in 'terra firme' and floodplain areas (Campos, Ribeiro, Souza-Junior, Ribeiro-Filho & Almeida, 2012), Cerrado fields (Martins, Ferreira, Curi, Vitorino & Silva, 2006), and meadows (Kubitski, 1979). Notwithstanding, although Humaitá and Borba have the largest number of species records among the cities south of the Madeira river, subsampling may occur due to the low proportion of angiosperm species per 10 km 2 ( Figure 5).
Collection gaps are large for angiosperms, but much larger for other groups such as bryophytes, ferns, lycophytes, gymnosperms, and fungi (Coelho & Amorim, unpublished data). In this sense, it is necessary to continuously create intensive projects to make local floras (Hopkins, 2007(Hopkins, , 2019Forzza et al., 2010;Lopes et al., 2014). These projects should ensure that new collection data are incorporated into herbariums so that the true biological diversity of these areas can be known and rare species unknown to science can be found (Ribeiro et al., 1994;Hopkins, 2007Hopkins, , 2019Forzza et al., 2010).

Conclusion
This study highlights the importance of floristic studies in the Brazilian Amazon. Given the well-known sample deficiency of the Amazon region, and considering the countless anthropogenic pressures that cities south of the Madeira river have been facing, this preliminary angiosperm checklist provides the first tool for further botanical studies in this region. The survey highlights the record of 47 new occurrences for Amazonas State. Of these new occurrences, 12 are also the first record for the northern region. In addition, a new species of vine of the genus Mandevilla Lindl was identified.