Introduction
Tropical mountains harbour a wide biological diversity, especially with regard to flora (Martinelli, Reference Martinelli2007). An important factor governing plant diversity in tropical mountains is the high heterogeneity of environments found in these regions, which leads to the occurrence of species with different adaptations in small areas (Ribeiro et al., Reference Ribeiro, Opazo and Scarano2007). This shows that high beta diversity (Whittaker et al., Reference Whittaker, Willis and Field2001) is an important component of such systems.
A characteristic formation occurring in the mountains of south-eastern Brazil is the campos rupestres (rocky grasslands). This vegetation occurs on rocky outcrops and shallow soils, on the tops of mountains, associated with quartzites and sandstones (Benites et al., Reference Benites, Schaefer, Simas and Santos2007). It consists of a mosaic of vegetation types, ranging from open areas covered by herbaceous vegetation to areas of dense vegetation mainly comprising shrubs and small trees (Benites et al., Reference Benites, Schaefer, Simas and Santos2007; Vasconcelos, Reference Vasconcelos2011; Conceição et al., Reference Conceição, Rapini, Carmo, Brito, Silva, Neves, Jacobi and Fernandes2016). The Serra Negra, a component of the Mantiqueira mountain complex, has a vegetation mosaic in which patches of rain forest and campos rupestres predominate in the landscape. A floristic survey of the Serra Negra region has demonstrated its high species richness on a regional scale, with 1020 phanerogamic species (Salimena et al., Reference Salimena, Matozinhos, Abreu, Souza, Ribeiro and Menini-Neto2013), 209 pteridophyte species (Souza et al., Reference Souza, Salino, Viana and Salimena2012) and 93 bryophyte species (Amorim, Reference Amorim2013) catalogued so far. Many of these species are considered rare or endemic, and some are endangered. Species new to science, in the process of description, and new botanical records for the state of Minas Gerais have also been found (Salimena et al., Reference Salimena, Matozinhos, Abreu, Souza, Ribeiro and Menini-Neto2013).
The area of occurrence of campos rupestres is still a matter of discussion. Most authors agree that campos rupestres occur in the south-east of Brazil, principally in the transition between the Cerrado and Atlantic Forest Domains, in the Espinhaço Range, and other isolated areas (Benites et al., Reference Benites, Schaefer, Simas and Santos2007; Vasconcelos, Reference Vasconcelos2011; Alves et al., Reference Alves, Silva, Oliveira and Medeiros2014; Fernandes, Reference Fernandes and Fernandes2016). However, there is no consensus about the occurrence of campos rupestres in areas of the Mantiqueira mountain complex, such as Serra Negra and Ibitipoca, both in Minas Gerais State. Although Salimena et al. (Reference Salimena, Matozinhos, Abreu, Souza, Ribeiro and Menini-Neto2013) considered the open vegetation in Serra Negra to be campos rupestres similar to those encountered at Ibitipoca, Alves et al. (Reference Alves, Silva, Oliveira and Medeiros2014) point out that a woody scrub replaces the open vegetation in most parts of the Serra Negra.
The woody scrub highlighted by Alves et al. (Reference Alves, Silva, Oliveira and Medeiros2014) in the Serra Negra remains a poorly sampled formation. At present, there is no published work focusing specifically on this formation. Therefore, a specific study on the appropriate methodology for sampling woody plants is necessary. In this paper, we present the first analysis of the structure and diversity of non-forest woody vegetation on quartzite soils in this region of the Mantiqueira mountain complex, Minas Gerais State, Brazil.
Materials and Methods
Study area
The Serra Negra is located between the municipalities of Rio Preto, Lima Duarte, Santa Barbara do Monte Verde and Olaria, at approximately 21°58′43.95′′S, 43°52′16.85′′W (Fig. 1), and is part of the Mantiqueira mountain complex (Valente et al., Reference Valente, Garcia, Salimena and Oliveira-Filho2011). The elevation is 900–1698 m, mean annual temperature is 19.3°C and annual mean rainfall is 1886 mm (EMATER, 2003). The climate is Cwb, according to Köppen's classification (Peel et al., Reference Peel, Finlayson and Mcmahon2007): humid mesothermal with dry and cold winters, and mild and humid summers.
Fig. 1. Location of Serra Negra in the Atlantic Forest Domain, between the municipalities of Rio Preto, Lima Duarte, Santa Barbara do Monte Verde and Olaria, Minas Gerais, in the south-eastern region of Brazil.
The landscape is mountainous, with steep slopes and valleys (Heilbron et al., Reference Heilbron, Tupinambá, Eirado, Ribeiro, Paciullo, Trouw, Valeriano, Junho, Roig, Nogueira, Medeiros, Rocha, Polonia, Silva and Toledo2000). The predominant geological formation of the area belongs to the Andrelândia Group, with dystrophic yellow oxisol, in addition to the presence of sand material (quartzite) with low natural fertility (Olszevski et al., Reference Olszevski, Costa, Fernandes-Filho and Costa2008; Oliveira & Marques-Neto, Reference Oliveira and Marques-Neto2014). The quartzites occur as two rock types in the region: coarse quartzite, with > 95% of quartz grains ranging from 3 to 8 mm (covering about 28 km2), and impure quartzite, that is, quartz associated with feldspar and traces of muscovite, with grains between 1 and 3 mm (covering about 10 km2) (Uagoda et al., Reference Uagoda, Avelar and Netto2011).
The region is within the domain of the Atlantic Forest, with a mosaic of vegetation types consisting of forest, shrub and anthropogenic physiognomies. Forest types are described according to the Brazilian vegetation classification (IBGE, 2012) as Alluvial, Montane and Upper Montane Dense Rain Forests (Valente et al., Reference Valente, Garcia, Salimena and Oliveira-Filho2011) and Semideciduous Seasonal Forests. The campo rupestre physiognomy occurs mainly on the ridges, associated with quartzites (Salimena et al., Reference Salimena, Matozinhos, Abreu, Souza, Ribeiro and Menini-Neto2013). The non-forest woody formations (Fig. 2) that are the target of this study have developed on sandy soils of quartzite origin (coarse quartzite), being distributed across the landscape in patches among other physiognomies. These formations have been described as ‘broadleaved scrubs’ by Salimena et al. (Reference Salimena, Matozinhos, Abreu, Souza, Ribeiro and Menini-Neto2013) following the classification system proposed by Oliveira-Filho (Reference Oliveira-Filho2009), which also corresponds to ‘scrub’ according to Eiten (Reference Eiten1979).
Fig. 2. Woody formations on quartzite soils in Serra Negra, Minas Gerais, Brazil. A, Vegetation mosaic with patches of rain forest and the woody formations occurring on quartzite soils. B, The woody formations that were the target of this study.
Sampling and data collection
The sampling methodology was proposed for woody vegetation in grasslands (wood savannas) and rock vegetation types of central and north-eastern Brazil (Felfili et al., Reference Felfili, Carvalho and Haidar2005). Using the maps provided by Uagoda et al. (Reference Uagoda, Avelar and Netto2011) and satellite images, 10 patches of non-forest woody formations were randomly selected. In each selected patch, a plot of 20 × 50 m was placed in the centre of the area, in order to avoid edge effects. The 10 plots together covered 1 ha. Field data were collected between June and September 2012.
All individuals with a diameter ≥ 3 cm at 30 cm above ground level (diameter at ground height) were measured. The height of these individuals was estimated by comparison with a graduated rod. A digital caliper was used to measure the stem diameter (diameter at ground height). Botanical material was collected for identification at species level. Identification of botanical material was made by reference to the literature, consultation with experts and comparison with material deposited in the herbarium Leopoldo Krieger (CESJ) (code following Thiers, continuously updated) of the Federal University of Juiz de Fora. The names of the species and botanical synonyms were confirmed by consulting the Flora do Brasil 2020 (Rio de Janeiro Botanical Garden, continuously updated), and the classification of botanical families followed APG III (Angiosperm Phylogeny Group, 2009).
Data analysis
For the structural analysis, the following phytosociological parameters of species were evaluated according to formulas described in Kent & Coker (Reference Kent and Coker1992): absolute and relative frequency (AF and RF, respectively); absolute and relative density (AD and RD, respectively); absolute and relative dominance (ADo and RDo, respectively); and importance value (IV), calculated as the sum of RF, RD and RDo. To analyse dominance in multistemmed individuals (individuals that had two or more stems branching below 30 cm), we summed the basal areas of each stem (Felfili et al., Reference Felfili, Carvalho and Haidar2005). The Shannon diversity index (H′) was used in the natural logarithmic base (e) to estimate species diversity. By using a logarithmic basis, this index is more suitable for communities in which species abundance is very uneven, balancing the weight of dominant species (Magurran, Reference Magurran2004). The evenness index (J) based on H′ was used to estimate the community uniformity. The analyses were performed using the PAST software, version 2.08 (Hammer, Reference Hammer2011).
Histograms of diameter distribution for the community were developed. The class intervals were defined by the approximation of Spiegel's formula (Felfili & Resende, Reference Felfili and Resende2003), resulting in class intervals of 2 cm. Logarithmic fit, its equation and coefficient of correlation (R 2) for each distribution, was determined. A histogram of height distribution for the community was also prepared, using the class interval of 0.5 m.
Results
Richness and diversity
In total, 1899 individuals were sampled, with a mean of 287.9 ± 50.9 individuals/plot, belonging to 30 families and 68 species (see Appendix). Six taxa were identified to genus level and three to family level; one was not identified. The family with the largest number of species was Myrtaceae (16 species), followed by Asteraceae and Melastomataceae (five species each) and Lauraceae and Primulaceae (four species each). Annonaceae, Ericaceae, Euphorbiaceae, Fabaceae, Hypericaceae, Lamiaceae, Rubiaceae and Sapindaceae had two species each. All other families were represented by a single species. The Shannon diversity index (H′) was 2.74 nats/individual, and evenness (J) was 0.65.
Phytosociology and vegetation structure
Table 1 shows the phytosociological parameters of the community. As observed in the diversity values (H′ and J), the results show a community with strong ecological dominance, in which > 50% of IV is concentrated only in five species: Eremanthus incanus (Less.) Less., Eremanthus erythropappus (DC.) MacLeish, Eugenia modesta DC., Byrsonima variabilis A.Juss. and Trembleya parviflora (D.Don) Cogn.
Table 1. Phytosociological parameters of species sampled in vegetation on quartzitic soils in Serra Negra, Rio Preto, Minas Gerais*
*Species ordered by decreasing importance value (IV = RD + RDo + RF).
AD, absolute density; ADo, absolute dominance; AF, absolute frequency; RD, relative density; RDo, relative dominance; RF, relative frequency.
As seen in Figure 3, most individuals are concentrated in the smaller classes of diameter, gradually decreasing in the bigger classes, in a reverse J pattern (logarithmic fit, y = –608ln(x) + 1253; R 2 = 0.751). The same pattern was observed when the biomass (sum of basal areas of individuals) was analysed by diameter class, showing that the major fraction of biomass is accumulated in smaller individuals (Fig. 4; logarithmic fit, y = –0.96ln(x) + 2.172; R 2 = 0.920).
Fig. 3. Frequency distribution in diameter classes of stems in vegetation on quartzite soils in Serra Negra, Rio Preto, Minas Gerais. Logarithmic fit, R 2 = 0.751.
Fig. 4. Sum of basal areas per diameter class for vegetation on quartzite soils in Serra Negra, Rio Preto, Minas Gerais. Logarithmic fit, R 2 = 0.920.
The distribution of individuals in height classes shows that the community is characterised by a large number of small individuals, with some emergents reaching a maximum height of 6 m (Fig. 5). It is not possible to identify the formation of different strata. Furthermore, the vegetation is not continuous, occurring more densely at some sites and sparsely at others, even at small scales (Fig. 2).
Fig. 5. Frequency distribution in height classes for vegetation on quartzite soils in Serra Negra, Rio Preto, Minas Gerais.
Another important aspect is the large number of multistemmed individuals (Table 2), with about 28% of individuals multistemmed at ground height (30 cm above ground), and 48.5% of species with at least one multistemmed individual. Of the species with highest density (more than 10 individuals), only one, Tibouchina fissinervia (Schrank & Mart. ex DC.) Cogn. (species name listed in Flora do Brasil 2020, Rio de Janeiro Botanical Garden, continuously updated; Tibouchina fissinervia Cogn. in the Plant List, 2013–), had no multistemmed individuals. Species with higher IV also had a high percentage of multistemmed individuals (Table 2).
Table 2. Number of individuals, percentage of multistemmed individuals and total stems for the community and principal populations* in vegetation on quartzitic soils in Serra Negra, Rio Preto, Minas Gerais
*Ordered by decreasing importance value (see Table 1).
Discussion
Richness and diversity
The composition of most representative families sampled is close to that reported for the general flora of Serra Negra (Salimena et al., Reference Salimena, Matozinhos, Abreu, Souza, Ribeiro and Menini-Neto2013), and also approximates the composition of forests in the region (Valente et al., Reference Valente, Garcia, Salimena and Oliveira-Filho2011). The Myrtaceae, Asteraceae, Melastomataceae, Lauraceae and Primulaceae families are recognised as important in forest formations at altitude in the Atlantic Forest Domain (Oliveira-Filho & Fontes, Reference Oliveira-Filho and Fontes2000). Therefore, in general, the floristic composition of woody vegetation on quartzite soils on Serra Negra tends to approach that of other woody formations in the region.
There have been no studies using similar methodology in similar physiognomies in the Atlantic Forest Domain, which prevents a direct comparison with this domain. However, the Shannon diversity and evenness values found were close to those found for the savanna (cerrado sensu stricto) area on quartzite soils in Piauí State, northern Brazil (H′, 2.75 nats/individual; J, 0.70), for which the same sampling methodology was applied (Lindoso et al., Reference Lindoso, Felfili, Costa and Castro2009). Species richness was also close to that determined by Lima et al. (Reference Lima, Pinto, Lenza and Pinto2010) for some areas of rocky ‘cerrado’ in central Brazil, using the same sampling, although the H′ values were slightly higher (H′, 3.09–3.65 nats/individual). These comparisons reveal a close structural affinity with the Brazilian savannas vegetation, which deserves future phytogeographical investigation.
Phytosociology and vegetation structure
The community showed a typical reverse J pattern of diameter class distribution. It can be inferred that the community has the potential to maintain its structure over time, with a good stock of young individuals able to occupy the places left by dead ones.
The two species of greatest ecological importance, Eremanthus incanus and E. erythropappus, which together comprise 33% of IV, are commonly known as ‘candeias’ and occur frequently in the rocky physiognomies of the Cerrado; they also occur in the Atlantic Forest Domain at altitudes above 700 m a.s.l. (Macleish, Reference Macleish1987).
Ecological dominance is common in plant communities in habitats with more severe environmental conditions in the Atlantic Forest Domain (Scarano, Reference Scarano2002). In these environments, locally abundant species often play an important role in the operation and maintenance of communities (Scarano et al., Reference Scarano, Duarte, Ribeiro, Rodrigues, Barcellos, Franco, Brulfert, Deléens and Lüttge2001). In the present study, the strong dominance contrasts with the high species richness recorded in the area (Salimena et al., Reference Salimena, Matozinhos, Abreu, Souza, Ribeiro and Menini-Neto2013). This is probably the result of limiting environmental factors, such as poor soil (Ribeiro, Reference Ribeiro2013), favouring species with adaptations to these conditions. Microhabitats in that environment and the proximity to other vegetation types (Menini Neto et al., Reference Menini Neto, Matozinhos, Abreu, Valente, Antunes, Souza, Viana and Salimena2009; Valente et al., Reference Valente, Garcia, Salimena and Oliveira-Filho2011) allow the establishment of a large number of local rare species. Some of the species found were among the most important species recorded in forest formations adjacent to the study areas (Valente et al., Reference Valente, Garcia, Salimena and Oliveira-Filho2011). Although these species have low densities and frequencies in campos rupestres, they are important species in forests in the Serra Negra. According to a study of forest formations at Serra Negra (Valente et al., Reference Valente, Garcia, Salimena and Oliveira-Filho2011), Alchornea triplinervia (Spreng.) Müll.Arg., for example, is the species with the highest IV in the Upper Montane Dense Rain Forest, and the species with the third highest IV in the Alluvial Dense Rain Forest; and Eugenia widgrenii Sond. ex O.Berg. and Xylopia brasiliensis Spreng. are the species with the second and third highest IV in the Montane Dense Rain Forest. This sharing of species between different vegetation types in the Serra Negra was also observed by Salimena et al. (Reference Salimena, Matozinhos, Abreu, Souza, Ribeiro and Menini-Neto2013). Sharing of species with forest formations seems to be a frequent pattern in grassland environments of mountains, where forest species tend to occur as occasional elements, showing smaller sizes, and often being more branched, because of limitations imposed by edaphic aspects (Messias et al., Reference Messias, Leite, Meira-Neto and Kozovits2012; Oliveira-Filho et al., Reference Oliveira-Filho, Fontes, Viana, Valente, Salimena, Ferreira, Forzza, Menini Neto, Salimena and Zappi2013; Valente et al., Reference Valente, Araújo, Fontes, Rocha, Forzza, Menini Neto, Salimena and Zappi2013).
The high percentage of multistemmed individuals shows that the formation of multiple stems is an important process in the structural conformation of this community. In open vegetation, competition for light is not an important factor, therefore plants can show more diverse forms, such as the formation of multistemmed individuals, in response to different selective pressures (Archibald & Bond, Reference Archibald and Bond2003). Investment in forming multiple stems occurs at the expense of investment in other processes, especially the allocation of resources to reproduction (Iwasa & Kubo, Reference Iwasa and Kubo1997). Furthermore, multistemmed plants tend to have lower biomass and shorter height than single-stemmed plants, being less competitive in shaded environments (Bond & Midgley, Reference Bond and Midgley2001). However, sprouting and regrowth ability may have advantages as a way to recover after disturbances capable of causing loss of above-ground biomass, such as fire, strong winds and stem cut (Bond & Midgley, Reference Bond and Midgley2001). Theoretical models show that sprouting is advantageous in situations in which the frequency and intensity of disturbance is intermediate to intense (Iwasa & Kubo, Reference Iwasa and Kubo1997; Bellingham & Sparrow, Reference Bellingham and Sparrow2000).
The occurrence of high proportions of multistemmed individuals is not related only to the occurrence of disturbances in the environment. Dunphy et al. (Reference Dunphy, Murphy and Lugo2000), studying tropical dry forests in Puerto Rico, found a high proportion of multistemmed individuals and showed that in most cases they showed no signs of having been cut, keeping the main stem intact. Indeed, the model proposed by Bellingham & Sparrow (Reference Bellingham and Sparrow2000) incorporates productivity of the environment as an important factor in predicting the rate of sprouting. In environments with low productivity, as in cases of low soil fertility, sprouting can be advantageous even without the presence of constant disturbances. Bellingham & Sparrow (Reference Bellingham and Sparrow2009) found that, in tropical montane forests in Jamaica, the formation of multistemmed individuals is mainly related to low soil fertility. In the present study, it is likely that the high proportion of multistemmed individuals, especially among the most abundant species, is also related to low soil fertility (Olszevski et al., Reference Olszevski, Costa, Fernandes-Filho and Costa2008).
Final considerations and implications for conservation
The results show that the woody vegetation on quartzite soils of Serra Negra is characterised by the contrast of strong ecological dominance and expressive richness, with elements of adjacent forests. In addition, three sampled species are considered endangered in the Atlantic Forest Domain of Minas Gerais State (Biodiversitas Foundation, 2008); Pseudobrickellia angustissima (Spreng. ex Baker) R.M.King & H.Rob. and Verbesina pseudoclaussenii D.J.N.Hind (Asteraceae) are considered critically endangered, and Handroanthus albus (Cham.) Mattos (Bignoniaceae) is considered vulnerable. According to the Biodiversitas Foundation (2008), which used the criteria of the International Union for Conservation of Nature, these three species are considered endangered mainly because of factors related to their area of occupation and loss of habitat. Martinelli & Moraes (Reference Martinelli and Moraes2013) also point out that Verbesina pseudoclaussenii is critically endangered in Brazil and is now restricted to the Serra Negra.
The two most abundant species, Eremanthus incanus and E. erythropappus, have great economic importance in the region because of their highly durable wood, which is used mainly for making fences and for firewood extraction. Information from local residents indicates that they commonly use the wood for making fences and rustic furniture, and as fuel for wood stoves, but still with a low impact on natural populations. In Brazil, the commercial exploitation of ‘candeias’ for extraction of essential oil, which is used in dermatological products and the cosmetics industry (Bohlmann et al., Reference Bohlmann, Zdero, King and Robinson1980; Silvério et al., Reference Silvério, Sousa, Del-Vechio-Vieira, Miranda, Matheus and Kaplan2008), has been growing. The imminent exploitation of these species for oil extraction may pose an additional risk to the vegetation studied here, because the results show that these species have great ecological importance in the structural conformation of the woody community.
The great heterogeneity of vegetation, with many species of low frequency and locally ‘rare’, means that a strategy for their effective conservation must include the largest possible area distributed across the landscape. Strategies for flora conservation in the region should focus not only on the rare and endangered species, but also on locally abundant species that are functionally vital in maintaining the plant community (Scarano, Reference Scarano2002) and should also be considered in biodiversity conservation plans (Scarano Reference Scarano2009). Commercial use of woody species, if not performed in a planned way, can also increase soil exposure, which exposes the fragile structure and makes it subject to erosion (Olszevski et al., Reference Olszevski, Costa, Fernandes-Filho and Costa2008). Therefore, a greater coordinated effort between the government and the local population is necessary to ensure the conservation and proper use of natural resources in the region.
Acknowledgements
We thank our colleagues at the herbarium Leopoldo Krieger (CESJ), Universidade Federal de Juiz de Fora, for their assistance in botanical identification; students of the Laboratory of Plant Ecology and Department of Botany, Universidade Federal de Juiz de Fora, especially Eduardo T. Amorim, for their help in fieldwork; researchers Dr João Marcelo Alvarenga Braga and Dr Luiz Menini Neto, and two anonymous reviewers, for their comments and suggestions; and the Postgraduate Programme in Ecology, Universidade Federal de Juiz de Fora, for logistical support. This work was supported by the Brazilian National Council of Technological and Scientific Development under grant no. 472921/2011-8 and master's scholarship for the first author.
Appendix
The testimony material (Appendix table 1) is deposited in the herbarium Leopoldo Krieger (CESJ) and the Laboratory of Plant Ecology, Department of Botany, Universidade Federal de Juiz de Fora.
Appendix table 1. Species sampled in vegetation on quartzitic soils in Serra Negra, Rio Preto, Minas Gerais, arranged in alphabetical order of family, genus and species, respectively
