Soil islands on rocky surfaces often harbor aggregated vegetation that consists of insular plant communities. These islands are typical of the rocky outcrops and in various parts of Brazil form the so-called "campos rupestres" vegetation. Four of such sites have been selected in the state of Bahia, Northeast Brazil, for this comparative study on floristics and vegetation structure: three areas situated inside the "Parque Nacional da Chapada Diamantina" (Guiné, Fumaça and "Gerais da Fumaça") and one is at the border of the Environmental Protection Area of "Marimbus-Iraquara" ("Mãe Inácia"). All occurring vegetation islands were studied in four random plots of 10 × 10 m per site. Soil was often shallow, sandy and acidic. Vascular plant species were determined, with respective life forms and canopy coverage areas. The total number of species when all four sites were added was 135, and the number of species per island varied from 2 to 32. The areas of the 214 soil islands varied from 0.015 to 91.9 m², totaling 568 m² in the four sites. Monocotyledon families were dominant, essentially Velloziaceae, as well as Orchidaceae, Bromeliaceae, Amaryllidaceae and Cyperaceae. Among the eudicotyledons, dominant families were mainly Clusiaceae, Asteraceae and Melastomataceae. The biological spectra revealed that phanerophytes and hemicryptophytes predominated among the life forms, while chamaephytes had the largest coverage area. Epilithic and desiccant chamaephytes composed the most conspicuous interspecific associations, and were probably related to early successional processes. Sites closest to one another were not the most similar in structure, indicating that other factors more relevant than distance might be involved in the abundance of species in space.
floristics; rocky outcrops; soil islands on rocks; vegetation islands; vegetation structure
Ilhas de solo abrigam comunidades vegetais agregadas no espaço e delimitadas pela superfície rochosa. Tais ilhas são típicas dos afloramentos rochosos de várias regiões brasileiras e, por vezes, integram parte da vegetação chamada de campo rupestre. Quatro sítios caracterizados por elevadas proporções de superfície rochosa foram selecionados para este estudo de florística e estrutura vegetacional: três situados dentro do Parque Nacional da Chapada Diamantina (Guiné, Fumaça e Gerais da Fumaça) e um no limite com a APA Marimbus-Iraquara (Mãe Inácia). Sortearam-se quatro parcelas de 10 × 10 m por sítio, onde todas as ilhas de solo ocorrentes tiveram suas áreas estimadas. O solo foi em geral raso, arenoso e ácido. Foram determinadas as espécies de plantas vasculares, com respectivas formas de vida e áreas de cobertura vegetal. O número de espécies nos quatro sítios foi 135, variando de 2 a 32 espécies por ilha. As áreas das 214 ilhas de solo variaram de 0,015 a 91,9 m², somando 568 m² nos quatro sítios. Famílias de monocotiledôneas foram as dominantes, especialmente Velloziaceae, além das Orchidaceae, Bromeliaceae, Amaryllidaceae e Cyperaceae. Dentre as eudicotiledôneas ressaltam-se as Clusiaceae, Asteraceae e Melastomataceae. Os espectros biológicos revelaram as formas de vida fanerófita e hemicriptófita como predominantes, apesar da caméfita possuir maior área de cobertura. Espécies camefíticas epilíticas e camefíticas dessecantes formaram as associações mais conspícuas, provavelmente relacionadas com os processos iniciais de sucessão da vegetação na rocha. Os sítios mais próximos entre si não foram os mais similares na estrutura, indicando outro(s) fator(es) mais relevante(s) do que a distância, envolvidos na abundância das espécies no espaço.
afloramentos rochosos; estrutura vegetacional; florística; ilhas de solo sobre rocha; ilhas de vegetação
Floristics, structure and soil of insular vegetation in four quartzite-sandstone outcrops of "Chapada Diamantina", Northeast Brazil
Florística, estrutura e solo da vegetação insular em quatro afloramentos de quartzito-arenito na Chapada Diamantina, nordeste do Brasil
Abel Augusto ConceiçãoI, 1 1 Corresponding author: firstname.lastname@example.org ; José Rubens PiraniII; Sergio Tadeu MeirellesIII
IUniversidade Estadual de Feira de Santana, Departamento de Ciências Biológicas, km 03 BR 116, Campus Universitário, 44031-460 Feira de Santana, BA, Brazil
IIUniversidade de São Paulo, Instituto de Biociências, Departamento de Botânica, Caixa Postal 11461, 05422-970 São Paulo, SP, Brazil
IIIUniversidade de São Paulo, Instituto de Biociências, Departamento de Ecologia Geral, Caixa Postal 11461, 05422-970 São Paulo, SP, Brazil
Soil islands on rocky surfaces often harbor aggregated vegetation that consists of insular plant communities. These islands are typical of the rocky outcrops and in various parts of Brazil form the so-called "campos rupestres" vegetation. Four of such sites have been selected in the state of Bahia, Northeast Brazil, for this comparative study on floristics and vegetation structure: three areas situated inside the "Parque Nacional da Chapada Diamantina" (Guiné, Fumaça and "Gerais da Fumaça") and one is at the border of the Environmental Protection Area of "Marimbus-Iraquara" ("Mãe Inácia"). All occurring vegetation islands were studied in four random plots of 10 × 10 m per site. Soil was often shallow, sandy and acidic. Vascular plant species were determined, with respective life forms and canopy coverage areas. The total number of species when all four sites were added was 135, and the number of species per island varied from 2 to 32. The areas of the 214 soil islands varied from 0.015 to 91.9 m2, totaling 568 m2 in the four sites. Monocotyledon families were dominant, essentially Velloziaceae, as well as Orchidaceae, Bromeliaceae, Amaryllidaceae and Cyperaceae. Among the eudicotyledons, dominant families were mainly Clusiaceae, Asteraceae and Melastomataceae. The biological spectra revealed that phanerophytes and hemicryptophytes predominated among the life forms, while chamaephytes had the largest coverage area. Epilithic and desiccant chamaephytes composed the most conspicuous interspecific associations, and were probably related to early successional processes. Sites closest to one another were not the most similar in structure, indicating that other factors more relevant than distance might be involved in the abundance of species in space.
Key words: floristics, rocky outcrops, soil islands on rocks, vegetation islands, vegetation structure
Ilhas de solo abrigam comunidades vegetais agregadas no espaço e delimitadas pela superfície rochosa. Tais ilhas são típicas dos afloramentos rochosos de várias regiões brasileiras e, por vezes, integram parte da vegetação chamada de campo rupestre. Quatro sítios caracterizados por elevadas proporções de superfície rochosa foram selecionados para este estudo de florística e estrutura vegetacional: três situados dentro do Parque Nacional da Chapada Diamantina (Guiné, Fumaça e Gerais da Fumaça) e um no limite com a APA Marimbus-Iraquara (Mãe Inácia). Sortearam-se quatro parcelas de 10 × 10 m por sítio, onde todas as ilhas de solo ocorrentes tiveram suas áreas estimadas. O solo foi em geral raso, arenoso e ácido. Foram determinadas as espécies de plantas vasculares, com respectivas formas de vida e áreas de cobertura vegetal. O número de espécies nos quatro sítios foi 135, variando de 2 a 32 espécies por ilha. As áreas das 214 ilhas de solo variaram de 0,015 a 91,9 m2, somando 568 m2 nos quatro sítios. Famílias de monocotiledôneas foram as dominantes, especialmente Velloziaceae, além das Orchidaceae, Bromeliaceae, Amaryllidaceae e Cyperaceae. Dentre as eudicotiledôneas ressaltam-se as Clusiaceae, Asteraceae e Melastomataceae. Os espectros biológicos revelaram as formas de vida fanerófita e hemicriptófita como predominantes, apesar da caméfita possuir maior área de cobertura. Espécies camefíticas epilíticas e camefíticas dessecantes formaram as associações mais conspícuas, provavelmente relacionadas com os processos iniciais de sucessão da vegetação na rocha. Os sítios mais próximos entre si não foram os mais similares na estrutura, indicando outro(s) fator(es) mais relevante(s) do que a distância, envolvidos na abundância das espécies no espaço.
Palavras-chave: afloramentos rochosos, estrutura vegetacional, florística, ilhas de solo sobre rocha, ilhas de vegetação
Rocky outcrop vegetation is often aggregated on soil islands surrounded by the exposed rock (McCormick et al. 1974, Wiser et al. 1996). Severe environmental conditions, such as soil scarcity and nutrient deficiency, high daily temperature oscillations and intense irradiation restrict the occupation of the exposed rock by plants (Shure & Ragsdale 1977, Burrows 1990, Ware 1990), although these unfavorable conditions are somewhat ameliorated when the vegetation is denser (Daubenmire 1968). These plant aggregations surrounded by rock surfaces are called "soil mat communities" (Hambler 1964), "island communities" (Burbanck & Platt 1964, McCormick et al. 1974, Shure & Ragsdale 1977), "vegetation islands" (Medina et al. 2006) or "soil islands" (Conceição & Pirani 2005). In this paper we adopt the latter terminology.
Studies on soil islands on rocks are more abundant in granitic-gneiss outcrops in Africa and in the United States. In South America, Ibisch et al. (1995) referred to families such as Bromeliaceae, Cactaceae, Cyperaceae, Orchidaceae, Poaceae, and Velloziaceae as typical of local inselbergs, while Velloziaceae is typical also of African inselbergs (Sarthou & Villiers 1998, Michelangeli 2000). In Brazil, some of the earliest studies on plant communities in rock outcrops were those of Segadas-Vianna (1965), Oliveira et al. (1975) and Carauta & Oliveira (1982). Again, Bromeliaceae, Velloziaceae and Orchidaceae appeared as important components in rocky slopes and/or high altitude zones of Rio de Janeiro, organized in "plant clumps that were similar to islands", as described by Carauta & Oliveira (1982).
In Brazil, there appears to be a very restricted number of taxa that are common on rocky outcrop vegetation (França et al. 1997, Meirelles et al. 1999, Ribeiro & Medina 2002, Conceição & Pirani 2005, 2007, Medina et al. 2006, Conceição et al. 2007). The "Chapada Diamantina" region, inserted in the semi-arid "caatinga" biome of the Brazilian northeast (Giulietti & Pirani 1988), has been chosen for this study for its high biodiversity, high level of endemisms and the presence of quartzite-sandstone rock outcrops (Harley & Simmons 1986, Giulietti et al. 1987, 1996, 1997, Giulietti & Pirani 1988, Alves & Kolbek 1994, Harley 1995, Stannard 1995, Conceição 2000, Conceição & Pirani 2005, 2007). According to Harley & Simmons (1986), "campo rupestre" is the expression that defines the vegetation that grows on quartzite-sandstone substrate, although this terminology is used to refer to Brazilian vegetation on other types of rocky substrate such as granitic-gneiss (e.g., Queiroz et al. 1996) or ferruginous (i.e., the locally called "canga", e.g., Viana & Lombardi 2007). Still according to Harley & Simmons (1986), "campos rupestres" occur in altitudes higher than 900 m in the Brazilian states of Goiás, and some isolated areas in São Paulo and Rio de Janeiro, and along the Espinhaço mountain chain in Minas Gerais and Bahia States.
"Morro do Pai Inácio", an inselberg with elevations between 1,100 and 1,170 m, is the single location at "Chapada Diamantina" that has been studied in respect to the vegetation ecology on soil islands (Conceição et al. 2007). This "campo rupestre" vegetation is composed essentially of shrubs and herbs (Conceição & Giulietti 2002). The present study adds four other locations to the list of study sites of "Chapada Diamantina" as regards the ecology of vegetation on soil islands. To each of them, we aimed to describe and discuss the relationships between edaphic conditions, floristics and vegetation structure.
Material and methods
Study sites This study was carried out at four "campo rupestre" sites with a great proportion of exposed rock on hilltops and ridge tops at "Serra do Sincorá" ("Sincorá" Range), "Chapada Diamantina", state of Bahia, Northeast Brazil (figure 1). "Mãe Inácia" site (12°27'S and 41°28'W) is located in the municipality of Palmeiras, and is the only site outside the "Parque Nacional da Chapada Diamantina", with elevations between 1,100 and 1,140 m a.s.l. It is the most isolated hill of our four study sites and has four small summits (1 to 3 ha) isolated from each other by three cleefs of 5 to 10 m of width. "Cachoeira da Fumaça" (12°35' S and 41°27'W) and "Gerais da Fumaça" (12°36' S and 41°28'W) outcrops are located at "Serra da Larguinha", in the municipality of Palmeiras, between 1,310 and 1,360 m of altitude. "Gerais da Fumaça" is the less isolated site from the surrounding vegetation. Guiné outcrops are found on the western border of "Serra do Sincorá" and "Parque Nacional da Chapada Diamantina", at "Serra do Esbarrancado", municipality of Mucugê (12°45' S and 41°30' W). They are the highest outcrops (ca. 1,400 m a.s.l.) among our study sites.
The climate at the study sites is Central Brazil Tropical, sub-hot, semi-humid, with a humid summer and four to five dry months concentrated during spring. From June to August, months are cooler and the first morning hours are often cloudy. Mean annual temperatures at locations with elevations between 1,000 and 1,100 m are lower than 20 °C, and minimum daily temperatures lower than 4 °C may occur (Nimer 1989). Predominant winds come from the southeast and orographic rains occur in the eastern sector, where our study sites are found (Jesus et al. 1985).
Sandstones and quartzites formed in the Pre-Cambrian make up the rugged topography of the "Chapada Diamantina", resulting from differential erosion (Moreira & Camelier 1977). The rocky outcrops studied are included in the "Chapada Diamantina" Group, which starts near Mucugê, extending through "Santo Inácio", and is part of the Tombador Formation (Torquato & Fogaça 1981). The soil sediments upon which the vegetation stands are shallow, sandy, and acidic, concentrating high organic matter and clay contents in relation to locations with more continuous vegetation (Conceição & Giulietti 2002, Conceição & Pirani 2005).
Soil islands The soil islands were taken as sample units defined as combinations of two or more individuals of vascular plants sharing the same patch of soil that is surrounded by a bare rock surface devoid of any vascular plants (Conceição & Pirani 2005). In each of the four sites, four random 10 × 10 m plots were drawn, and within these limits all soil islands were labeled and numbered. Islands only partly included within the plots were also sampled entirely. Thus, establishing plots was a strategy only to standardize the islands sampled.
Soil sampling The collection of soil samples was hindered by the small quantity of sediments in the soil islands, both at surface and in depth. Shallow sediments appeared associated to fractures and other unevenness on the rock surfaces. The sediments were composed mainly by organic and inorganic material and root mass, while ant and termite nests were often found.
For each site, two out of the four 10 × 10 m plots were randomly chosen for the extraction of two sediment samples at 10 cm of depth in each of them, totaling four soil samples per site. These collections were made in the central portions of the largest and the smallest island that had at least 10 cm of depth. In one of the plots at "Morro da Mãe Inácia", only one island had 10 cm of depth, which meant a smaller number of samples for this site (15 soil samples rather than 16).
Soil samples were analyzed at the Department of Soils and Plant Nutrition at the "Escola Superior de Agricultura Luiz de Queiroz", Piracicaba, state of São Paulo, following the methods described by Ruggiero et al. (2002). Between-site comparison of edaphic parameters were done by one-way ANOVA (Callegari-Jacques 2003), calculated by the software Statistica 6.0. Initially, variables were tested for normality (Kolmogorov-Smirnov, P > 0.05) and homocedasticity (Levene, P > 0.05). The hypothesis tested by ANOVA was that means were equal in each site (H0: µM = µF = µGF = µG, a = 0.05; where sites are M = "Mãe Inácia", F = Fumaça, GF = "Gerais da Fumaça" e G = Guiné). Rejection of H0 meant between-site difference, and Tukey test (a = 0.05) was applied to test such differences (Callegari-Jacques 2003).
Surface area of vegetation islands and cover area of individual species The surface areas of the islands were estimated by measuring their longest axis (l) and greatest width (w). Then, based on the general form of the island, either the area of an ellipse (¼ l.w.p) or a rectangle (l.w) was calculated. The percent area covered by the different plant species composing each island was visually estimated based on the vertical projection of all aerial plants parts (Westhoff & Maarel 1978).
Species composition and life forms Fertile specimens of all species on the soil islands were collected and oven-dried. Voucher specimens were deposited at the herbarium SPF ("Universidade de São Paulo"). Family names followed Cronquist (1981), except for Amaryllidaceae and Fabaceae s.l. (= Leguminosae). The abbreviations of the names of the botanical authorities followed Brummitt & Powell (1992). Raunkiaer's life-forms were determined according to Ellenberg & Müeller-Dombois (1967) and further details are provided in Conceição & Pirani (2005).
Data analyses Community structure was explored applying the multivariate statistical techniques of classification and ordination on the patterns of species distribution and abundance in soil islands of the different sites.
A cluster procedure was used for representation of species co-occurring in soil islands on all sites. The procedure was applied on a Jaccard similarity coefficient matrix calculated from the species presence-absence data on soil islands. Only species that occurred in more than 9% of the soil islands were included in the analysis. The construction of the similarity dendrogram was based on the unweighed pair-group mean clustering algorithm (UPGMA).
For exploring the patterns arising from species abundance by site, an ordination procedure by correspondence analysis (CA) of the detrended variant (DCA) was applied as a strategy for unbiased indirect gradient detection (Ter Braak 1997). This ordination procedure was applied on a matrix of sites and species abundance values in soil islands occurring on 100 m2 plots. Detrending was done by segments. The ordination of sites on the two first axes was examined by interpolation of DCA scores resulting from the DCA.
Detrended correspondence analysis ordinations and the r-mode UPGMA cluster analyses were performed using MVSP version 3.12 software (Kovach Computing Service, Anglesey, UK) (Kovach 1999).
Soil Soil analyses showed marked variations in regard to the edaphic features of our four study sites (table 1). Calcium concentration was higher at Fumaça, as compared to "Mãe Inácia" and "Gerais da Fumaça" (FCa(3, 11) = 5.9066, P = 0.0118). The percentage of silt was higher also at Fumaça as compared to "Gerais da Fumaça" (Fsilt(3, 11) = 5.5784, P = 0.0142). Aluminum concentration was higher at "Mãe Inácia" than at Fumaça and Guiné (F (3, 11) = 9.8631, P = 0.0019). At Fumaça, pH was more acidic than at "Mãe Inácia" (F (3, 11) = 6.7698, P = 0.0075). However, three out of seven edaphic variables that satisfied the assumptions of ANOVA were statistically similar between the study sites (P > 0.05): organic matter (F (3, 11) = 1.0783), cation exchange capacity (F (3, 11) = 0.8456) and sodium (F (3,11) = 0.2444).
Vegetation Table 2 lists the 135 species of angiosperms (distributed in 43 families) found in the 214 soil islands of four rocky outcrops of "Chapada Diamantina". There were nine species included in seven families of vascular cryptogams. Eudycotyledons were represented by 58 species included in 24 families. The family Piperaceae was represented by one species and monocots were 67 species of 11 families. The surface area of the soil islands varied from 0.015 to 91.9 m2, and 2 to 32 plant species were found per island. The total surface area of soil islands when all 214 islands of our four sites were added was 568 m2, while the total coverage of vascular plant species in these islands was 675 m2. The difference between the two values is due to overlaps between branches of different species in given soil islands. Thus, there was an average of 0.23 plant species m-2 of soil islands. The site Fumaça had the highest number of species (85) on 216 m2 of soil islands (0.39 species m-2), while "Gerais da Fumaça" had the lowest (39 species m-2) on 213 m2 of soil islands (0.18 species m-2). Guiné had 57 species on 74 m2 (0.76 species m-2) and "Morro da Mãe Inácia" had 42 on 63 m2 (0.66 species m-2) of soil islands.
Orchidaceae was the most species-rich family (26), followed by Poaceae (12), Asteraceae (10), Velloziaceae and Bromeliaceae (seven each), Cyperaceae and Fabaceae (five each), and Verbenaceae and Melastomataceae (four each). Most families (27; 63%) had only one or two species. Out of the ten most frequent families, seven were monocots (figure 2A). The surface area covered by species belonging to the Velloziaceae family amounted to 43% of the 675 m2 of total coverage area when all vascular species belonging to the 214 soil islands are added (figure 2B).
Figure 3 shows percentage occurrence (A) and percent cover (B) of the seven most representative botanical families per site. From all families in the two legends, six were the same throughout (Velloziaceae, Asteraceae, Cyperaceae, Bromeliaceae, Orchidaceae, and Melastomataceae), while Amaryllidaceae was representative for occurrence and Clusiaceae for aerial plant coverage. "Gerais da Fumaça" had the highest proportion (47%) of occurrence of species belonging to "other families", i.e., those that do not belong to the seven plant families in the legend of figure 3A, whereas in "Morro da Mãe Inácia" these families were the most representative (80%). In terms of aerial coverage, the seven families were again more strongly represented in the latter site (figure 3B).
Phanerophyte and hemicryptophyte life forms were predominant, while therophytes were numerous in site "Gerais da Fumaça" (figure 4A). Percent cover was expressive for desiccation-tolerant chamaephyte, phanerophyte and thallo-chamaephyte (lichens), while hemicryptophytes covered high proportions of sites "Gerais da Fumaça" and Guiné, and epilithic chamaephyte did so at "Mãe Inácia" (figure 4B).
Classification and ordination analysis The classification of the 22 species according to their occurrence in 214 soil islands revealed distinct patterns of floristic composition among the soil islands. The dendrogram in figure 5 shows that there are four distinct groups of co-occurring species in soil islands. These groups are related to specific sites, which was evidenced by the presence of exclusive species for "Mãe Inácia" (Barbacenia blanchetii and Vellozia hemisphaerica), Fumaça (Vellozia punctulata and Vellozia dasypus), Guiné (Orthophytum albopictum) and "Gerais da Fumaça" and Guiné (Vellozia jolyi). In addition to the exclusive species, there were exclusive combinations of species for the soil islands, such as the pair Trilepis lhotzkiana and Cattleya elongata in the "Mãe Inácia" site. The islands characterized by the combination of Orthophytum albopictum and Mandevilla bahiensis occurred at Guiné, and included Lasiolaena duartei and Vriesea atra. Islands containing the combination of Hippeastrum solandriflorum, Abildgaardia sp., Schizachyrium sanguineum and Tibouchina pereirae were typical for "Gerais da Fumaça" and Guiné, but are common species in soil islands of all sites.
The two first axes resulting from DCA analysis of the abundance matrix represented together 31.64% of the total variation in the matrix with 23.22% and 8.42% in the first and the second axis respectively. A gradient can be depicted through the interpolation of the site scores on the first two axes. The arrangement of site scores is polarized by the "Mãe Inácia" plots in opposition to Fumaça, while Gerais and "Gerais da Fumaça" are in intermediate positions. The second axis discriminates the Gerais and "Gerais da Fumaça" plots. All the plots appear in very distinct groups in the interpolation of the two first axes scores. The "Mãe Inácia" plots occur in a small gradient on the first axis, which suggests a higher heterogeneity in this site.
Soil type and site isolation The soil type of the vegetation islands and the degree of isolation of the study sites were often two of the main factors related to between-site variation in diversity and species composition for the four sites studied at "Chapada Diamantina". This pattern probably results in a high beta-diversity, as already described for inselbergs in the state of Rio de Janeiro (Meirelles et al. 1999).
The shallow, sandy and acidic soil of the vegetation islands of "Chapada Diamantina" resembled those of other "campos rupestres" (Duarte 1967, Joly 1970, Harley 1995, Vitta 1995, Conceição & Giulietti 2002, Conceição & Pirani 2005). Since community structure is influenced by resource availability (Huggett 1995), soil type is likely to be related to the differences in diversity between sites found here. For instance, the high percentages of silt and base saturation, plus the high concentration of phosphorus and low aluminum saturation, might help explain the highest species richness found at the Fumaça site. The higher humidity in this site is due to a neighboring waterfall that might benefit decomposers (Huggett 1995), as well as the plants directly. Moreover, the finest soil particles (clay) are related to a higher capacity of nutrient retention (Raven et al. 2001). Conversely, the "Gerais da Fumaça" site that had the lowest species richness also had low percentages of silt, clay, organic matter, low capacity for cation exchange, soil acidity, and a relatively high concentration of aluminum. The "Morro da Mãe Inácia" site had similarly low species richness, and despite a high concentration of organic matter, all other parameters resembled "Gerais da Fumaça". Guiné, which was the site with intermediate species richness, also had intermediate values for soil parameters, particularly for organic matter, base saturation and aluminum.
Species composition of "Morro da Mãe Inácia", the most isolated site, revealed a number of microendemisms, not found on the other sites less isolated from the surrounding vegetation. However, although this is an indication that isolation should affect species composition and local endemisms, proximity between sites did not grant species composition similarity. For instance, "Gerais da Fumaça" and Fumaça were not the most similar to each other. This was surprising since these sites were closest to each other and were both adjacent to extensive fields, which might have favored in both cases the entrance of grassland species. This is an indication that factors other than distance alone might be more relevant to determine spatial patterns of species occurrence, such as soil discussed above.
Floristic patterns Despite the regional variation between sites and the overall diversity found, the taxonomic entities of the studied flora largely resemble those of other rock outcrops in Brazil. For instance, the predominance of monocots confirms a well-known pattern for neotropical rock outcrops (Ibisch et al. 1995, Meirelles 1996, Porembski et al. 1998, Meirelles et al. 1999, Gröger 2000, Safford & Martinelli 2000, and Ribeiro et al. 2007, in this issue). The main families of vascular plants found here were also found elsewhere in "Chapada Diamantina" (Conceição et al. 2007) and even in south-east Brazil (e.g., Velloziaceae, Bromeliaceae, Orchidaceae, Cyperaceae, Asteraceae, Amaryllidaceae, Poaceae, Melastomataceae, Clusiaceae). Interestingly, there are some species co-occurrences with nepheline-syenite outcrops at high altitude in Itatiaia, Rio de Janeiro State (Ilex amara, Peperomia galioides and Polypodium catharinae; Ribeiro & Medina 2002) and in granitic outcrops in Atibaia, São Paulo State (four vascular cryptogams, Tillandsia stricta, Trilepis lhotzkiana, Hillia parasitica and Epidendrum secundum [syn. Epidendrum elongatum Lindl.]; Meirelles 1996, Meirelles et al. 1999).
Similarly, genera such as Abildgaardia, Acianthera (Pleurothallis), Barbacenia, Clusia, Epidendrum, Hippeastrum, Lychnophora, Orthophytum, Polypodium, Tibouchina, Trilepis, Vellozia and Vriesea are commonly found in our study site and in other rock outcrops in Brazil (Oliveira et al. 1975, Carauta & Oliveira 1982, Ibisch et al. 1995, Meirelles 1996, França et al. 1997, Porembski et al. 1998, Waldemar 1998, Meirelles et al. 1999, Safford & Martinelli 2000, Conceição et al. 2007). Although some of these genera are broadly distributed, many species are endemic to "Chapada Diamantina" including new species found that belong to the genera Barbacenia (Velloziaceae), Paepalanthus (Eriocaulaceae) and Jacquemontia (Convolvulaceae) which confers a regional floristic peculiarity.
Life-forms Life-form abundance and distribution patterns also resemble that found elsewhere in Brazil: 1) the predominance of phanerophytes and hemicryptophytes (such as in Meirelles 1996, Meirelles et al. 1999, Ribeiro & Medina 2002); 2) the high cover area of desiccation-tolerant chamaephytes in locations with a high proportion of exposed rocks (see also Meirelles et al. 1997, 1999, Conceição & Giulietti 2002, Conceição & Pirani 2005, Conceição et al. 2007); 3) the relatively high proportion of cryptophytes (e.g., Mandevilla bahiensis, Hippeastrum spp., and several orchids) suggest an adaptive value for underground structures capable of tolerance to fire, drought (see also Conceição 2003) and trampling by cattle; 4) the commonness of thallo-chamaephytes demonstrates the importance of lichens in such vegetation type (see also Burrows 1990). However, the high proportion of therophytes found at "Gerais da Fumaça" is less commonly described in other neotropical rock outcrops and might result from the proximity with surrounding grassland fields, as already mentioned, which harbor many such species prone to temporary waterlogging.
"Morro do Pai Inácio" (surveyed by Conceição et al. 2007) had a higher number of species per area than the sites studied here (table 3). Interestingly, this site had the lowest mean island size (total island size/number of islands). This seems to suggest that, considering two cases where total island areas are the same, the one with a high number of small islands is likely to harbor more species than that with a low number of larger islands. This pattern points to the environmental heterogeneity and competition as probable sources of species richness in these rock outcrops. Similar patterns have been described elsewhere (Virolainen et al. 1998).
The well defined groups obtained from the classification analysis of soil islands can be related to the occurrence of defined island "types", similar in composition and structure. These island types can be represented by species combinations more or less related to the site of occurrence. The cluster analysis presented some of these "island types" in groups. Some examples were the Trilepis lhotzkiana Cattleya elongata and Vellozia hemisphaerica Acianthera ochreata islands. These islands types are dominated by plants with desiccation tolerant behavior (T. lhotzkiana and V. hemisphaerica) being probably related to primary succession patterns on "Mãe Inácia" site. Similar compositions were found on Fumaça with combinations of Vellozia punctulata V. dasypus and Epidendrum secundum. Clues on the cause of the clear segregation of plot scores resulting from DCA (figure 6), can be found on the difference in surface roughness (crested on "Mãe Inácia" and smooth on Fumaça.) and site specific features such as available humidity on Fumaça site, due to the proximity of a waterfall.
One of the most remarkable features of rock outcrop vegetation is the tendency to harbor microendemisms that can be produced even on a scale in which is unlike to occur segregation. This pattern is possibly related to the combination of the rock outcrop extreme conditions and species competitive advantages and requirements in the limited space of the soil islands. Soil, as well as minor local microclimate variations, can be responsible for a higher competitive advantage of a given species on a particular island. Thus, the observed effect is the arrangement of typical species assemblages on soil islands on rock outcrops situated nearby as well as microendemisms. Such features emphasize the extreme sensitivity of rock outcrop plant populations as the microendemics can become extinct with minor environmental changes produced by human interference or natural dynamics of the substrate and microclimate.
Acknowledgements We thank IBAMA for research permits; CAPES and FAPESP (99/05322-7) for grants for the first author; F.R. Scarano for valuable comments and translation advice; and the following plant specialists: A. Giulietti (Eriocaulaceae), A. Grillo (Fabaceae s.l.), A. Rapini (Asclepiadaceae), A. Zanin (Poaceae), A.C. Araújo (Cyperaceae), A.L. Toscano-de-Brito (Orchidaceae), A. Martins (Melastomataceae), C. Garcia (Poaceae), C. Mônica (Araceae), C. van den Berg (Orchidaceae), D. Zappi (Cactaceae), E. Jacques (Begoniaceae), E. Smidt (Orchidaceae), F.N. Costa (Eriocaulaceae), F.R. Salimena (Verbenaceae), F. Vitta (Cyperaceae), G. Pedralli (Dioscoreaceae), H. Bautista (Asteraceae), H. Longhi-Wagner (Poaceae), I. Cordeiro (Euphorbiaceae), J. Baumgratz (Melastomataceae), J. Coffani-Nunes (Bromeliaceae), J. Prado (criptógamas vasculares), J. Semir (Asteraceae), J. Wurdack (Melastomataceae), L. Funch (Myrtaceae), L. Queiroz (Fabaceae s.l.), L.R. Lima (Euphorbiaceae), M. Alves (Cyperaceae), M. Groppo (Aquifoliaceae), M. Paciencia (criptógamas vasculares), M. Wanderley (Xyridaceae), N. Roque (Asteraceae), N.D. Hind (Asteraceae), P. Labiak (criptógamas vasculares), P. Sano (Eriocaulaceae), R. Forzza (Bromeliaceae), R. Harley (Lamiaceae), R. Mello-Silva (Velloziaceae), R. Simão-Bianchini (Convolvulaceae), S. Atkins (Verbenaceae), T. Silva (Verbenaceae) and W. W. Thomas (Cyperaceae).
(received: August 31, 2004; accepted: December 6, 2007)
- ALVES, R.J.V. & KOLBEK, J. 1994. Plant species endemism in savanna vegetation on table mountains (campo rupestre) in Brazil. Vegetatio 113:125-139.
- BRUMMITT, R.K. & POWELL, C.E. 1992. Authors of plant names. Royal Botanic Gardens, Kew.
- BURBANCK, M.P. & PLATT, R.B. 1964. Granite outcrop comunities of the Piedmont Plateau in Georgia. Ecology 45:292-306.
- BURROWS, C.J. 1990. Processes of vegetation change. Urwin Hyman, London.
- CALLEGARI-JACQUES, S.M. 2003. Bioestatística: princípios e aplicações. Artmed, Porto Alegre.
- CARAUTA, J.P. & OLIVEIRA, R.R. 1982. Fitogeografia das encostas do Pão de Açúcar. Alguns estudos II. Série trabalhos técnicos nş 2/82. Fundação Estadual de Engenharia do Meio Ambiente. Secretaria de Estado de Obras e Serviços Públicos, Rio de Janeiro.
- CONCEIÇÃO, A.A. 2000. Alerta para a conservação da biota na Chapada Diamantina. Ciência Hoje 27:54-56.
- CONCEIÇÃO, A.A. 2003. Ecologia da vegetação em afloramentos rochosos na Chapada Diamantina, Bahia, Brasil. Tese de doutorado, Universidade de São Paulo, São Paulo.
- CONCEIÇÃO, A.A. & GIULIETTI, A.M. 2002. COMPosição florística e aspectos estruturais de campo rupestre em dois platôs do Morro do Pai Inácio, Chapada Diamantina, Bahia, Brasil. Hoehnea 29:37-48.
- CONCEIÇÃO, A.A. & PIRANI, J.R. 2005. DELImitação de habitats em campos rupestres na Chapada Diamantina: substratos, composição florística e aspectos estruturais. Boletim de Botânica da Universidade de São Paulo 23: 85-111.
- CONCEIÇÃO, A.A. & PIRANI, J.R. 2007. Diversidade em quatro áreas de campos rupestres na Chapada Diamantina, Bahia, Brasil: espécies distintas, mas riquezas similares. Rodriguésia 58:193-206.
- CONCEIÇÃO, A.A., GIULIETTI, A.M. & MEIRELLES, S.T. 2007. Ilhas de vegetação em afloramentos de quartzito-arenito no Morro do Pai Inácio, Chapada Diamantina, Bahia, Brasil. Acta Botanica Brasilica 21: 335-347.
- CRONQUIST, A. 1981. An integrated system of classification of flowering plants. Columbia University Press, New York.
- DAUBENMIRE, R.F. 1968. Plant communities: a textbook of plant synecology. Harper & Row, London.
- DUARTE, A.C. 1967. Alguns aspectos geográficos do Planalto da Conquista e da Chapada Diamantina. Boletim Geográfico 26:39-65.
- ELLENBERG, H. & MÜLLER-DOMBOIS, D. 1967. Tentative physiognomic-ecological classification of plant formations of the earth. Bericht des Geobotanischen Instituts der E.T.H., Stiftung Rübel in Zürich 37:21-55.
- FRANÇA, F., MELO, E. & SANTOS, C.C. 1997. Flora de inselbergs da região de Milagres, Bahia: caracterização da vegetação e lista de espécies de dois inselbergs, Brasil. Sitientibus 17:163-184.
- GIULIETTI, A.M., MENEZES, N.L., PIRANI, J.R., MEGURO, M. & WANDERLEY, M.G.L. 1987. Flora da Serra do Cipó, Minas Gerais: caracterização e lista das espécies. Boletim de Botânica da Universidade de São Paulo 9:1151.
- GIULIETTI, A.M. & PIRANI, J.R. 1988. Patterns of geographic distribution of some plant species from the Espinhaço Range, Minas Gerais and Bahia, Brazil. In Proceedings of a workshop on neotropical distribution patterns (P.E. Vanzolini & W.R. Heyer, eds.). Academia Brasileira de Ciências, Rio de Janeiro, p.39-69.
- GIULIETTI, A.M., PIRANI, J.R. & HARLEY, R.M. 1997. Espinhaço Range Region, Eastern Brazil. In Centres of plant diversity. A guide and strategy for their conservation. v.3. The Americas (S.D. Davis, V.H. Heywood, O. Herrera-Macbryde, J. Villa-Lobos & A.C. Hamilton, eds.). IUCN Publication Unity, Cambridge, p.397-404.
- GIULIETTI, A.M., QUEIROZ, L.P. & HARLEY, R.M. 1996. Vegetação e flora da Chapada Diamantina, Bahia. Anais da 4Ş Reunião Especial da SBPC, Feira de Santana, Bahia, p.144-156.
- GRÖGER, A. 2000. Flora and vegetation of inselbergs of Venezuelan Guayana. In Inselbergs (S. Porembski & W. Barthlott, eds.). Ecological Studies, v.146. Springer-Verlag, Berlin, p.291-314.
- GUEDES, M.L.S. & ORGE, M.D.R. 1998. Check-list das espécies vasculares do Morro do Pai Inácio (Palmeiras) e da Serra da Chapadinha (Lençóis), Chapada Diamantina, Bahia, Brasil. Universidade Federal da Bahia, Salvador.
- HAMBLER, D.J. 1964. The vegetation of granitic outcrops in western Nigeria. Journal of Ecology. 52:573-594.
- HARLEY, R.M. 1995. Introduction. In Flora of the Pico das Almas, Chapada Diamantina, Brazil (B.L. Stannard, ed.). Royal Botanic Gardens, Kew, p.1-42.
- HARLEY, R.M. & SIMMONS, N.A. 1986. Florula of Mucugê, Chapada Diamantina Bahia, Brazil. Royal Botanic Gardens, Kew.
- HUGGETT, R.J. 1995. Geoecology: an evolutionary approach. Routledge, New York.
- IBISCH, P.L., RAUER, G., RUDOLPH, D & BARTHLOTT, W. 1995. Floristic, biogeographical, and vegetational aspects of Pre-Cambrian rock outcrops (inselbergs) in eastern Bolivia. Flora 190:299-314.
- JESUS, E.F.R., FALK, F.H., RIBEIRO, L.P. & MARQUES, T.M. 1985. Caracterização geográfica e aspectos geológicos da Chapada Diamantina, Bahia. Centro editorial e didático da Universidade Federal da Bahia, Salvador.
- JOLY, A.B. 1970. Conheça a vegetação brasileira. Edusp & Polígono, São Paulo.
- KLUGE, M. & BRULFERT, J. 2000. Ecophysiology of vascular plants on inselbergs. In Inselbergs (S. Porembski & W. Barthlott, eds.). Ecological Studies, v.146. Springer-Verlag, Berlin, p.143-173.
- KOVACH, W.L. 1999. MVSP A Multivariate Statistical Package for Windows, Kovach Computing Services, Wales.
- MCCORMICK, J.F., ARIEL, E.L. & SHARITZ, R.R. 1974. Experimental analysis of ecosystems In Vegetation and environment (B.R. Strain & W.D. Billings, eds.). Dr. W. Junk, The Hague, p.151-179.
- MEDINA, B.M.O., RIBEIRO, K.T. & SCARANO, F.R. 2006. Plant-plant and plant-topography interactions on a rock outcrop at high altitude in southeastern Brazil. Biotropica 38:27-34.
- MEIRELLES, S.T. 1996. Estrutura da comunidade e características funcionais dos componentes da vegetação de um afloramento rochoso. Tese de doutorado, Universidade Estadual de Campinas.
- MEIRELLES, S.T., MATTOS, E.A. & SILVA, A.C. 1997. Potential desiccation tolerant vascular plants from Southeastern Brazil. Polish Journal of Environmental Studies 6:17-21.
- MEIRELLES, S.T., PIVELLO, V.R. & JOLY, C.A. 1999. The vegetation of granite rock outcrops in Rio de Janeiro, Brazil, and the need for its protection. Environmental Conservation 26:10-20.
- MICHELANGELI, F.A. 2000. Species composition and species-area relationships in vegetation isolates on summit of a sandstone mountain in southern Venezuela. Journal of Tropical Ecology 16:69-82.
- MOREIRA, A.A.N. & CAMELIER, C. 1977. Relevo. Geografia do Brasil. IBGE, Rio de Janeiro.
- NIMER, N.1989. Climatologia do Brasil. IBGE, Rio de Janeiro.
- OLIVEIRA, R.F., COIMBRA FILHO, A.F. & SILVA, Z.L. 1975. Sobre litosere: algumas espécies para revestimento de encostas rochosas. Brasil Florestal 6:3-18.
- POREMBSKI, S., MARTINELLI, G., OHLEMÜLLER, R. & BARTHLOTT, W. 1998. Diversity and ecology of saxicolous vegetation mats on inselbergs in the Brazilian atlantic rainforest. Diversity and Distribution 4:107-119.
- POREMBSKI, S., SEINE, R. & BARTHLOTT, W. 2000. Factors controlling species richness of inselbregs. In Inselbergs (S. Porembski & W. Barthlott, eds.). Ecological Studies, v.146. Springer-Verlag, Berlin, p.451-481.
- QUEIROZ, L.P., SENA, T.S.N. & COSTA, M.J.S.L. 1996. Flora vascular da Serra da Jibóia, Santa Terezinha Bahia. I: o campo rupestre. Sitientibus 15:27-40.
- RAVEN, P.H., EVERT, R.F. & EICHHORN, S.E. 2001. Biologia vegetal. Guanabara Koogan, Rio de Janeiro.
- RIBEIRO, K.T. & MEDINA, B.M.O. 2002. Estrutura, dinâmica e biogeografia das ilhas de vegetação sobre rocha do Planalto do Itatiaia, RJ. Boletim do Parque Nacional do Itatiaia 10:11-82.
- RUGGIERO, P.G.C., BATALHA, M.A., PIVELLO, V.R. & MEIRELLES, S.T. 2002. Soil-vegetation relationships in cerrado (Brazilian savanna) and semideciduous forest, Southeastern Brazil. Plant Ecology 160:1-16.
- SAFFORD, H.D. & MARTINELLI, G. 2000. Southeast Brazil. In Inselbergs (S. Porembski & W. Barthlott, eds.). Ecological Studies, v.146. Springer-Verlag, Berlin, p.339-389.
- SARTHOU, C. & VILLIERS, J. 1998. Epilithic plant communities on inselbergs in French Guiana. Journal of Vegetation Science 9:847-860.
- SEGADAS-VIANNA, F. 1965. Ecology of the Itatiaia range, southeastern Brazil. I Altitudinal zonation of the vegetation. Arquivos do Museu Nacional 53:7-30.
- SHURE, D.J. & RAGSDALE, L. 1977. Patterns of primary succession on granite outcrop surfaces. Ecology 58:993-1006.
- STANNARD, B.L. 1995. Flora of the Pico das Almas, Chapada Diamantina, Brazil. Royal Botanic Gardens, Kew.
- TER BRAAK, C.F.J. 1997. Ordination. In Data analysis in community and landscape ecology. (R.H.G. Jongman, C.J.F. Ter Braak & O.F.R. van Tongeren, eds.). Cambridge University Press, New York.
- TORQUATO, J.R. & FOGAÇA, A.C.C. 1981. Correlação entre o Supergrupo Espinhaço no Brasil, o Grupo Chela em Angola e as Formações Nasib e Khoabendus da Namíbia. In Anais do simpósio sobre o Craton do São Francisco e suas faixas marginais. Sociedade Brasileira de Geologia Núcleo da Bahia, Coordenação da Produção Mineral, Salvador, p.87-99.
- VIANA, P.L. & LOMBARDI, J.A. 2007. Florística e caracterização dos campos rupestres sobre canga na Serra da Calçada, Minas Gerais, Brasil. Rodriguésia 58:159-177.
- VIROLAINEN, K.M., SUOMI, T., SUHONEM, J. & KUITUNEM, M. 1998. Conservation of vascular plants in single large and several small mires: species richness, rarity and taxonomic diversity. Journal of Applied Ecology 35:700-707.
- VITTA, F.A. 1995. Composição florística e ecologia de comunidades campestres na Serra do Cipó, Minas Gerais. Dissertação de mestrado, Universidade de São Paulo, São Paulo.
- WALDEMAR, C.C. 1998. A vegetação rupestre heliófila do Parque Estadual de Itapuã, Viamão-RS. Dissertação de mestrado, Universidade Federal do Rio Grande do Sul, Porto Alegre.
- WARE, S. 1990. Adaptation to substrate-and lack of it-in rock outcrop plants: Sedum and Arenaria American Journal of Botany 77:1095-1100.
- WISER, S.K., PEET, R.K. & WHITE, P.S. 1996. High-elevation rock outcrop vegetation of the Southern Appalachian Mountains. Journal of Vegetation Science 7:703-722.
Publication in this collection
25 Feb 2008
Date of issue
06 Dec 2007
31 Aug 2004