Life-forms, pollination and seed dispersal syndromes in plant communities on ironstone outcrops, SE Brazil

Formas de vida, síndromes de polinização e dispersão de sementes em comunidades vegetais sobre afloramentos ferruginosos, SE do Brasil

ARTICLES ARTIGOS

Life-forms, pollination and seed dispersal syndromes in plant communities on ironstone outcrops, SE Brazil

Formas de vida, síndromes de polinização e dispersão de sementes em comunidades vegetais sobre afloramentos ferruginosos, SE do Brasil

Claudia Maria Jacobi¹; Flávio Fonseca do Carmo

Universidade Federal de Minas Gerais, Instituto de Ciências Biológicas, Departamento de Biologia Geral, Belo Horizonte, MG, Brazil

RESUMO

Afloramentos rochosos têm um papel importante na diversidade vegetal de ecossistemas montanos. As cangas (afloramentos ferruginosos) estão entre os litotipos menos conhecidos e mais ameaçados do sudeste do Brasil, devido às atividades minerarias. Além da composição de espécies, um aspecto fundamental para promover sua conservação e restauração é o conhecimento das formas de vida, síndromes de polinização e dispersão de sementes dominantes. As análises foram baseadas em listas florísticas publicadas de cangas do sudeste do Brasil. Um total de 353 espécies de angiospermas (70 famílias) foi distribuído entre as duas fisionomias predominantes (áreas abertas e capões de mata) em cangas. Dezesseis famílias foram responsáveis por 70% do total de espécies. Comparado ao espectro normal de Raunkiaer, fanerófitos estiveram super-representados e terófitos sub-representados. Os primeiros foram a forma de vida predominante em capões, enquanto que os hemicriptófitos o foram em áreas abertas. A entomofilia foi a síndrome de polinização dominante em ambos os hábitats. A zoocoria foi dominante em capões e foi última em áreas abertas, onde a anemocoria e autocoria prevaleceram. Considerando que ambas as fisionomias estão sujeitas às mesmas condições climáticas, os resultados corroboram a influência de componentes geoedáficos nos três atributos analisados.

Palavras-chave: Afloramentos rochosos, cangas, conservação, condições geoedáficas, espectro biológico

ABSTRACT

Rock outcrops play an important role in enhancing plant diversity in montane ecosystems. Ironstone outcrops (cangas) are among the lithotypes less known and most threatened in SE Brazil, due to mining activities. Besides species composition, a key feature to promote their conservation and restoration is the knowledge of the community prevalent life-forms, pollination and seed dispersal syndromes. The analyses were done based on published floristic surveys of cangas in SE Brazil. A total of 353 species of angiosperms (70 families) were assigned to one of the two predominant physiognomies (open areas and forest islands) on ironstone outcrops. Sixteen families responded for 70% of all species. Compared to Raunkiaer's spectrum, phanerophytes were over- and therophytes were under-represented. Phanerophytes were the predominant life-form in forest islands, while hemicryptophytes were outstanding in open areas. Entomophily was the dominant pollination syndrome in both habitats. Zoochory was dominant in forest islands and ranked last in open areas, where anemochory and autochory prevailed. Considering that both forest islands and open areas are subjected to the same climatic conditions, the results corroborate the influence of geoedaphic components in the three traits analysed.

Key words: Biological spectrum, cangas, conservation, geoedaphic conditions, rock outcrops

Introduction

Mountain ranges are key landscape features promoting biological diversity (Burke 2003), due to unique factors related primarily to edapho-climatic variations resulting from altitudinal gradients, allied to the natural topographic discontinuity among mountaintops. Within these, rock outcrops have proven to substantially contribute to regional plant diversity. Over the past two decades the diversityenhancing role of rock outcrops on mountain environments in tropical and temperate regions has been confi rmed for some of the most important rock types, particularly granite, sandstone and ironstone (Barthlott et al. 1993; Giulietti et al. 1997; Porembski & Barthlott 2000; Kruckberg 2002; Jacobi et al. 2007; Jacobi & Carmo 2008a, b).

In tropical regions these rocky environments are in practice xeric islands within matrices of mesic vegetation (Szarzynski 2000), as a consequence of being edaphically controlled. Outcrops are usually characterized by very shallow -if any- soil cover except for depressions and crevices where organic matter is more abundant and humidity long-lasting (Porembski & Barthlott, 2000). Th ese edaphic conditions, allied to altitudinal climatic factors, promote a vast list of environmental stressors, ranging from absence of mechanical support to high daily temperature ranges and strong winds (causing high evapotranspiration and overheated bare substrates), high UV exposure, and low water retention (Saff ord & Martinelli 2000; Szarzynski 2000).

In Brazilian rock outcrops a considerable amount of information on plant communities is available among the well-studied granitoid inselbergs -gneiss, granite and nepheline-syenite domes (e.g. Meirelles et al. 1999; Ribeiro & Medina 2002; Caiafa & Silva 2007) and sandstones outcrops -quartzite and arenite table-mountains (e.g. Alves & Kolbek 1994; Meguro et al. 1994; Pirani et al. 2003; Conceição & Pirani 2005). Historically, much less is known about the vegetation associated with ironstones outcrops (cangas), considering that Brazil is among the few countries where these formations occur. In southeastern Brazil, the "Quadrilátero Ferrífero" (i.e., Iron Quadrangle) is a key economic area recognized as a major worldwide iron producer (IBRAM 2008; 2009), and one of the few in the country to harbor ironstone outcrops. These consist of a crust that covers the main iron ore deposits in the country. Even though they represent the predominant rock outcrop lithotype in the Iron Quadrangle, the flora associated with ironstones was practically ignored until recently, when increasing mining activity raised attention to their high plant diversity and the need for their protection (Jacobi et al. 2007, 2008; Jacobi & Carmo 2008b).

Although scanty, floristic surveys have already shown that ironstone outcrops contribute substantially to plant diversity in the Iron Quadrangle. Moreover, the species composition differs significantly from that of the surrounding matrix, which falls within the domain of the two Brazilian hotspots, Cerrado and Atlantic Forest (Jacobi & Carmo 2008a). Considering that cangas share the same environmental stresses and disjoint distribution of other rock outcrop types, but that they are subjected to immediate degradation threats and large habitat loss, it is important to broaden our knowledge of their associated plant community structure and dynamics, of which life-forms and dispersal strategies are respectively key features. Th erefore, a further step for their conservation and restoration is to determine which life forms are prevalent in these environments, and which strategies for pollination and seed dispersal are most successful. Our work was directed by three questions: (1) What is the floristic biological spectrum of ironstone-outcrop plant communities? (2) What habitat conditions does this biological spectrum reflect? (3) Which pollination and seed dispersal syndromes occur and in what proportion? With this in mind, the aims of this study were to characterize and discuss the life-forms, pollination and seed dispersal syndromes in plant communities of the two major ironstone outcrop physiognomies occurring in the Iron Quadrangle.

Material and methods

Ironstone outcrop vegetation data

The Iron Quadrangle (IQ), with an area of approximately 7200 km², is located in southeastern Brazil. It is the south end portion of the Espinhaço Range, which runs N-S and has a maximum altitude of ca. 2072m. This mountain range, with an estimated number of 4000 plant species, is considered one of the world centers of plant diversity (Giulietti et al. 1997). The climate is tropical sub-humid and the IQ region, in spite of a mean annual precipitation of 1500-1900 mm, may be subjected to water deficit of five to seven months (April-October) during the winter (Nimer & Brandão 1989). Temperatures close to 70ºC measured above patches of bare rock are not uncommon (F. F. Carmo, pers. obs.).

Our analysis of life forms and syndromes was based on the three current publications which have dealt with floristic surveys of ironstone outcrops (known as cangas) in SE Brazil for at least one year: Jacobi et al. (2007), Viana & Lombardi (2007) and Mourão & Stehmann (2007). All three studies were performed within the Iron Quadrangle (Fig. 1). Th e first was a study of ironstone outcrop vegetation in Serra do Curral (20º03'33' S - 44º00'34' W, 1460 m asl) and the southern sector of Serra da Moeda (20º20'06' S - 43º56'16' W, 1560 m asl) encompassing 45 ha, which rendered a list of 234 species of vascular plants (Jacobi et al. 2007). Th e second was undertaken in the northern sector of Serra da Moeda (20º05'35' S - 43º59'01' W, 900-1426 m asl) and complemented with herbarium specimens, totaling 246 species of angiosperms in 75 ha (Viana & Lombardi 2007). The last surveyed two sites within Serra do Tamanduá (19º53'08' S - 43º26'11' W, 845m asl; 19º51'06' S - 43º22'35' W, 1063m asl), and found 117 species of angiosperms in 35 ha (Mourão & Stehmann 2007).

Voucher specimens are cited in two of the published works (Mourão & Stehmann 2007; Viana & Lombardi 2007) and those of the species listed in Jacobi et al. (2007) are deposited in the same institution as the fi rst two, the herbarium of the Federal University of Minas Gerais (BHCB). In the final combined list, undetermined species (64 spp.), Pteridophytes sensu lato (12 spp.), and species common to the four sites were removed. Life-forms and syndromes of each species were assigned by examining voucher specimens, through extensive use of literature which included community-level studies (e.g. Buzato et al. 2000; Griz & Machado 2001; Batalha & Martins 2002; Freitas & Sazima 2006; Kinoshita et al. 2006; Martins & Batalha 2006; Conceição et al. 2007), and from direct fi eld observations, particularly those of Jacobi et al. (2007).

Although an extensive array of habitat types occur on ironstone outcrops (Jacobi et al. 2007) these were grouped in only two major vegetation physiognomies for analysis: open herbaceous-shrub vegetation and low forest-shrub islands, following Viana & Lombardi (2007). Open herbaceous-shrub vegetation (hereafter open vegetation or open areas) is the most characteristic of rupicolous communities, and plants grow over bare substrate or colonize small crevices (Fig. 2); water may be scarce even during the rainy season due to the thin layer of soil, so limiting nutrients and water stress are common (Vincent & Meguro 2008). Epilithic species, dwarf shrubs, branched and prostrate forms, and a variety of herbs dominate (Jacobi et al., 2007), and it is in this habitat that adaptations typical of xeric environments, such as sclerophylly and rehydration (notably Trilepis lhotzkiana Nees and genus Vellozia Vand.), are most commonly found. Patches of low forest-shrub vegetation (hereafter forest islands) thrive only in depressions, large crevices, and other structures which allow organic matter to accumulate, so plants have apparently less water and nutrient restrictions, forming thick patches of associated treelets, epiphytes, and lianas. Although there is a small overlap (<10%) in species occurring in both habitats, we assigned each to its most frequent habitat (^Appendix).

Life-forms

The normal biological spectrum, or phyto-climatic spectrum, proposed by Raunkiaer was compared to the frequency of life-forms of the whole outcrop community, as well as to each of the two main vegetation physiognomies separately. This spectrum acts as a null model and ratio deviations from it are used to provide an indication of whether a habitat is subjected to an unfavourable season (Batalha & Martins 2002), such as a dry period, in this context. Th e comparison of both physiognomies followed Raunkiaer´s classifi cation modified by Mueller-Dombois and Ellenberg (1974), to include lianas, epiphytes, and hemiparasites.

Syndromes

Each species was assigned to its primary pollination or seed dispersal syndrome. Pollination syndromes (Faegri & van der Pijl 1979) considered were by wind, bees, fl ies, moths and butterflies, insects (all other groups), birds, bats, and ambophily (pollination by both wind and insects). Th e primary seed dispersal syndromes were grouped in three categories (van der Pijl 1982; Howe & Smallwood 1982): anemochory, autochory and zoochory. Within these, autochory included both ballistic and gravity-assisted dispersal, while zoochory included epizoochory and endozoochory (external and internal passive carrying of diaspores), as well as synzoochory (deliberate transport of diaspores, represented by two species of Loranthaceae and one of Santalaceae).

Frequencies of life-forms between the two main vegetation physiognomies, and with the Raunkiaer proportions were compared with a G-test (goodnessof-fit, based on maximum-likelihood ratios) available in software BioEstat 3.0 (Ayres et al. 2003). No attempts were made to quantify either syndromes or life forms by species abundance (cover or density), since the analyses were based on floristic checklists.

Results and discussion

The combination of all the checklists of ironstone outcrops totaled 353 species of angiosperms, belonging to 70 families, in approximately 155 ha. Less than 4% of the species occurred in all four sites, and 64% in only one outcrop (^Appendix). The open vegetation physiognomy dominated in terms of area (ca. 70%) and 248 species were associated with it, while 105 species occurred in forest islands. Herbs dominated with almost 37% of all species, followed by shrubs and sub-shrubs. As expected, the percentage of trees and lianas was low, and mainly associated with forest islands.

In open areas the great majority of species were nanophanerophytes (< 2 m tall), some exceptions being Pseudobrickellia brasiliensis (Asteraceae), Agarista oleifolia (Ericaceae), Hyptis lippioides (Lamiaceae), Cinnamomum quadrangulum (Lauraceae), Solanum lycocarpum (Solanaceae), and Lippia hermannioides (Verbenaceae), none attaining 5 m, and therefore classified as microphanerophytes (2-8 m). The latter category composed the bulk of forest-island species, and the presence of a few species under 2 m seems to be related to patches of poor microenvironmental conditions rather than to life-form.

Six species are allegedly endemic to ironstone outcrops in the IQ (Viana 2008), and all occurred exclusively in open areas: Arthrocereus glaziovii (Cactaceae), Dyckia consimilis and Vriesea minarum (Bromeliaceae), Mimosa calodendron (Fabaceae), Sinningia rupicola (Gesneriaceae), and Oncidium gracile (Orchidaceae). All are chamaephytes except for S. rupicola (cryptophyte) and M. calodendron (phanerophyte). Bromeliads and S. rupicola are bird-pollinated, whereas the other three are entomophilous. Seed dispersal is by wind except in A. glaziovii (zoochory) and M. calodendron (barochory).

Sixteen families accounted for 70% (n = 248) of all species, therefore influencing the outcome of life-form and syndrome proportions (Table 1). It is common to fi nd prevalence of a particular dispersal mode or life-form within a family or genus because of the phylogenetic correlation, and this provides additional information on strategies that are most successful in different environments (Arbeláez & Parrado-Rosselli 2005). Of these families, ten were represented in the list by predominantly (> 50%) or exclusively phanerophytes. The three dispersal modes were evenly distributed among the 16 families, but wind-dispersal accounted for 42% of the species, a reflection of the high species richness of Asteraceae, Orchidaceae, and Apocynaceae, whereas autochory ranked last with 24%. The six families which are primarily animal-dispersed occur mainly in forest islands. Regarding pollination syndromes, entomophily prevailed in a broad sense, ranging from non-specialized flowers pollinated by several insect families, such as Asteraceae, Orchidaceae, and Verbenaceae, to those exclusively melittophilous (bee-pollinated), frequently bearing special anther openings (e.g., Melastomataceae and Solanum) or offering uncommon resources (e.g., oil in Malpighiaceae), which demand specific foraging skills or body structures. Grasses and sedges were exclusively wind-pollinated.

Overall, considering open areas and forest islands together, phanerophytes (mostly semideciduous) predominated (44%), followed by hemicryptophytes (26%) and chamaephytes (15%). The contribution of each physiognomy to this proportion (Fig. 3), however, was signifi cantly different. In forest islands, phanerophytes (microphanerophytes) predominated (71%) and all other life forms were poorly represented except for lianas (12%). Th e most representative genera with phanerophytic species were Solanum, Myrcia, Eugenia, Eremanthus, Miconia, Vismia, Guapira, and Myrsine. On the other hand, in open areas, phanerophytes (mostly sclerophyllous nanophanerophytes) and graminoid hemicryptophytes had similar representation of approximately one-third, followed by chamaephytes (19%). In this physiognomy the most representative genera with phanerophytic species were Baccharis, Mimosa, and Lippia, whereas Panicum, Paspalum and Rhynchospora were important hemicryptophytes.

Th e floristic biological spectrum of the entire canga vegetation (species from both open vegetation and forest-island physiognomies) is signifi cantly different from Raunkiaer's (or normal) biological spectrum (G-test = 12.19, d.f. = 4, p = 0.016), with phanerophytes over- and therophytes underrepresented, respectively. The scarcity of annual plants has already been noted for inselbergs from southeastern Brazil (Safford & Martinelli 2000), as well as cerrado vegetation compared to other savannas (Batalha & Martins 2002). When analyzed separately, the deviation of forest islands from the normal spectrum is noticeable (G-test = 47.98, d.f. = 4, p < 0.0001), the highest weight in the deviation corresponded to two life-forms: phanerophytes, with almost double the proportion of the normal spectrum (46%), and therophytes, with no species compared to the normal 13%. Comparatively, open areas did not contribute as extensively as forest islands to the deviation observed in the overall result (G-test = 12.28, d.f. = 4, p = 0.018). In agreement, the difference between life-form spectra was signifi cant also between both physiognomies, either when only the fi ve typical Raunkiaer categories were used (G-Test = 53.45, d.f. = 4, p < 0.0001) or when were they were redistributed to single out epiphytes, parasites, and lianas (G-Test = 56.31, d.f. = 7, p < 0.0001).

Life-forms reflect the predominant environmental conditions to which plant communities are subjected (Mueller-Dombois & Ellenberg 1974; Mera et al. 1999). Raunkiaer´s spectrum is useful to associate deviations of specifi c spectra to more favourable or unfavorable phytoclimatic conditions (Cain 1950). Although its use in tropical communities has been criticized because the principle that the level of protection of meristems/buds are related to low winter temperatures would not be relevant in the tropics (Sarmiento & Monasterio 1983; Rizzini 1997), rainfall regime stands as the unfavorable climatic factor that applies to tropical ecosystems. However, the comparison of open areas with forest islands, in practice subjected to the same climatic conditions, showed an impor-tant influence of the substrate in edaphic and microclimatic conditions. The same amount of rain falling in both habitats is retained for longer periods in forest islands, supported by a deeper litter layer, and allowing the establishment of treelets in detriment to herbs. Ironstone outcrops are not compact structures, and it is common to find cavities of diff erent sizes underneath what appears to be a hard crust (Simmons 1963; Jacobi et al. 2007), where nutrient and humidity conditions create a milder environment than on the surface. Roots may reach these cavities through fissures, and may therefore fi nd a more suitable refuge for their development. Th us, geoedaphic variables seem to be a key component determining life-form spectra in these plant communities.

The use of syndromes to determine the eff ective dispersal and pollination vector has also been subjected to criticism, since this analysis usually excludes potential dispersers and pollinators which may be of importance in certain situations (Howe & Smallwood 1982; Hingston & McQuillan 2000). Particularly in the case of pollination, associating each plant species to only one pollination syndrome must have masked the well-known generalist ability of floral visitors (Waser et al. 1996). Similarly, Griz & Machado (2001) also alerted to generalizations regarding dispersal of individual species, although they advocate this approach at the community level. Since this study did not analyze syndromes by predominance of individuals, our analysis cannot evaluate the success of each syndrome or species in terms of density or biomass. Nevertheless, they are still used as organizing tools, because they provide general answers regarding major or successful dispersal strategies at the community level.

Entomophily was the primary syndrome of 77% of all species in Table 1. Although it was predominant in both habitats, it accounted for more than 91% in forest islands, and only 70% in open areas (Fig. 4), where the second and third most representative categories were wind (16%) and hummingbird (13%) pollination. Among those pollinated by insects, melittophily prevailed in both physiognomies, and most of the outstanding bee-pollinated families were also common to both: Asteraceae, Fabaceae, Melastomataceae, Myrtaceae, and Solanaceae. In open areas, Convolvulaceae, Euphorbiaceae, Orchidaceae, and Velloziaceae were also important bee-pollinated families.

Syndromes of pollination by moths and butterfl ies were more frequent in forest islands, and cantarophily (beetle-pollination) occurred exclusively in these habitats, represented by two species of Guatteria Ruiz & Pav. Fly-pollination prevailed in open areas, associated with Acianthera Scheidw., Hydrocotyle L. and Aristolochia L. The majority (78%) of bird-pollinated species occurred in open areas, highlighting Bromeliaceae, Gesneriaceae, and Velloziaceae as the most frequent. In the entire community only four species were either ambophilous (Peperomia Ruiz & Pav.) or bat pollinated (Bauhinia rufa (Bong.) Steud. and Lafoensia pacari A.St.-Hil.).

Animal-dispersal of diaspores, primarily by frugivory, grouped 37% of all species in Table 1, followed closely by anemochory (33%) and autochory, with 30%. Comparing physiognomies, zoochory (mainly ornithochory) was predominant in forest islands (82%), but ranked last in open areas (18%), where autochory and anemochory (both with 41%) dominated (Fig. 5). Zoochory in forest islands was predominant in Myrtaceae, Solanaceae, Nyctaginaceae, Melastomataceae, and Rubiaceae. In open habitats the zoochoric families with more species were Solanaceae, Melastomataceae, and Bromeliaceae. The proportion of zoochory seems to be related to substrate properties, which influence the development of the root system and the ability to retain and store water. Among these outcrops, zoochory in forest islands, mainly in the form of dispersal by birds, reached a value that is more characteristic of tropical rainforests or at least ecosystems with a short dry season (Howe & Smallwood 1982; Safford & Martinelli, 2000).

Families with a higher number of wind-dispersed species in open habitats were Asteraceae, Orchidaceae, Apocynaceae, Poaceae, and Malpighiaceae and, in forest islands, Asteraceae, Bromeliaceae, Bignoniaceae, Orchidaceae, and Sapindaceae. Autochorous families in open habitats with more species were Poaceae, Cyperaceae, Velloziaceae, Verbenaceae, Melastomataceae (all of these predominantly gravity-assisted), Fabaceae, and Euphorbiaceae (the last two with some representatives of ballistic dispersal). In forest islands the families with more self-dispersed species were Fabaceae and Acanthaceae.

Animal dispersal of both seeds and pollen was found in 36% of the species, and 51% showed syndromes that indicate animal vectors for one of them, highlighting the importance of mutualistic relations in the maintenance of population processes. Only 13% of species, on the other hand, showed both abiotic pollination and seed dispersal. Of the species that presented biotic syndromes of both pollination and dispersal, 71% belonged to forest islands. On the other hand, 95% of the species that exploit abiotic pollination and dispersal vectors belonged to herbaceous vegetation. As expected, anemochorous species predominated in the open environment, but there was a surprisingly high percentage of zoophilous species in this physiognomy, particularly insect-pollinated.

Some syndromes were strongly associated with certain life forms, such as bat pollination with phanerophytes, entomophily with hemiparasites, lianas and phanerophytes, and anemochory with therophytes and hemicryptophytes. Considering the clear contrast between the biological spectra of open areas and forest islands, we can observe an association between habitats, life forms and pollination syndromes, a phenomenon corroborated by Ramírez (2004) in the Venezuelan savannas. In the entire community, bat-pollination was rare and observed only in phanerophytes. Ambophily occurred in chamaephytes (2%), and hemicryptophytes (1%). Wind pollination syndrome was observed in therophytes (59%), hemicryptophytes (38%), and less so in chamaephytes (2%). Bird-pollination was more frequent among cryptophytes (45%) and chamaephytes (27%).

Insect-pollination was the only syndrome occurring in all life forms, including hemiparasites (100%), lianas (93%) and phanerophytes (92%). Overall, zoochory was not observed in cryptophytes, but occurred in hemiparasites (Loranthaceae and Santalaceae), most phanerophytes (59%), and lianas (37%). Wind-dispersal was prevalent in therophytes (75%), cryptophytes (73%), and lianas (53%). Autochory was more frequent among hemicryptophytes (47%), chamaephytes (35%) and cryptophytes (27%).

In spite of being subjected to the same climate, the life-form spectra and syndromes in forest islands and open herbaceous vegetation are considerably different, and overall they respond to both habit (e.g. tree-shrubs vs. herbs) and physiognomy (forest islands vs. herbs-shrubs), which in turn are the product of geoedaphic conditions and their relationship with other environmental variables. Forest islands on ironstone outcrops had profiles poorly related with outcrop ecosystems, recognized as open systems; rather, they showed closer affi nity with the surrounding Cerrado and Atlantic Forest matrix physiognomies, underlining the milder micro-climatic conditions to which they are subjected. One aspect in common is that species in both physiognomies appear to be highly dependent on animals, especially pollinators, for their continued existence, and it is possible that this group might decline in abundance or diversity in a near future, considering the anthropic pressures to which cangas are subjected. How this might affect the structure and dynamics of these ecosystems is a topic that demands future studies, if possible integrating quantitative analyses of the plant communities, their vectors and substrate characteristics.

Acknowledgments

The authors thank FAPEMIG (Proc. APQ387/08) and CNPq (Proc. 479834/2008-3) for financial support. Th e comments of two anonymous reviewers greatly improved an early version of the manuscript.

References

Alves, R.J.V. & Kolbek, J. 1994. Plant species endemism in savanna vegetation on table mountains (Campo Rupestre) in Brazil. Vegetatio 113: 125-139.

Arbeláez, M.V. & Parrado-Roselli, A. 2005. Seed dispersal modes of the sandstone plateau vegetation of the Middle Caquetá River region, Colombian Amazonia. Biotropica 37: 64-72.

Ayres, M.; Ayres Jr., M.; Ayres, D.L. & Santos, A.S. 2003. BioEstat 3.0. Aplicações estatísticas nas áreas das ciências biológicas e médicas. Belém, Sociedade Civil Mamirauá .

Barthlott, W.; Groger, A. & Porembski, S. 1993. Some remarks on the vegetation of tropical inselbergs: diversity and ecological diff erentiation. Biogeographica 69: 105-124.

Batalha, M.A. & Martins, F.R. 2002. Life-form spectra of Brazilian cerrado sites. Flora 197: 452-460.

Burke, A. 2003. Inselbergs in a changing world - global trends. Diversity and Distributions 9: 375-383. Buzato, S.; Sazima, M. & Sazima, I. 2000. Hummingbird-pollinated floras at three Atlantic Forest sites. Biotropica 32: 824-841.

Caiafa, A.N. & Silva, A.F. 2007. Structural analysis of the vegetation on a highland granitic rock outcrop in Southeast Brazil. Revista Brasileira de Botânica 30: 657-664.

Cain, S.A. 1950. Life-forms and phytoclimate. Botanical Review 16: 1-32.

Conceição, A.A. & Pirani, J.R. 2005. Delimitação de habitats em campos rupestres na Chapada Diamantina, Bahia: substratos, composição florística e aspectos estruturais. Boletim de Botânica da Universidade de São Paulo 23: 85-111.

Conceição, A.A.; Funch, L.S. & Pirani, J.R. 2007. Reproductive phenology, pollination and seed dispersal syndromes on sandstone outcrop vegetation in the "Chapada Diamantina", northeastern Brazil: population and community analyses. Revista Brasileira de Botânica 30: 475-485. Faegri, K. & van der Pijl , L. 1979. The principles of pollination ecology. Oxford, Pergamon Press.

Freitas, L. & Sazima, M. 2006. Pollination biology in a tropical high-altitude grassland in Brazil: interactions at the community level. Annals of the Missouri Botanical Garden 93: 465-516.

Giulietti, A.M.; Pirani, J.R. & Harley, R.M. 1997. Espinhaço Range region - Eastern Brazil. Pp. 397-404. In: Davis, S.D. Heywood, V.H. Herrera-MacBryde, O. Villa-Lobos, J. & Hamilton, A.C. (Eds.). Centres of plant diversity: a guide and strategy for their conservation. vol. 3. The Americas. Cambridge, WWF/IUCN Publications Unit.

Griz, L.M.S. & Machado, I.C.S. 2001. Fruiting phenology and seed dispersal syndromes in caatinga, a tropical dry forest in the northeast of Brazil. Journal of Tropical Ecology 17: 303-321.

Hingston, A.B. & McQuillan, P.B. 2000. Are pollination syndromes useful predictors of floral visitors in Tasmania? Australian Journal of Ecology 25: 600-609.

Howe, H.F. & Smallwood, J. 1982. Ecology of seed dispersal. Annual Review of Ecology and Systematics 13: 201-228.

IBRAM - Instituto Brasileiro de Mineração. 2008. Informações e análises da economia mineral brasileira. Disponível em: www.ibram.org

IBRAM - Instituto Brasileiro de Mineração. 2009. A indústria de mineração em Minas Gerais. Disponível em: www.ibram.org

Jacobi, C.M.; Carmo, F.F.; Vincent, R.C. & Stehmann, J.R. 2007. Plant communities on ironstone outcrops - a diverse and endangered Brazilian ecosystem. Biodiversity and Conservation 16: 2185-2200.

Jacobi, C.M. & Carmo, F.F. 2008a. Diversidade dos campos rupestres ferruginosos no Quadrilátero Ferrífero, MG. Megadiversidade 4: 25-33.

Jacobi, C.M. & Carmo, F.F. 2008b. The contribution of ironstone outcrops to plant diversity in the Iron Quadrangle, a threatened Brazilian landscape. Ambio 37: 324-326.

Jacobi, C.M., Carmo, F.F. & Vincent, R.C. 2008. Vegetação sobre canga e seu potencial para reabilitação ambiental no Quadrilátero Ferrífero, MG. Revista Árvore 32: 345-353.

Kinoshita, L.S.; Torres, R.B.; Forni-Martins, E.R.; Spinelli, T.; Ahn, Y.J. & Constâncio, S.S. 2006. Composição florística e síndromes de polinização e de dispersão da mata do Sítio São Francisco, Campinas, SP, Brasil. Acta Botanica Brasilica 20: 313-327.

Kruckberg, A.R. 2002. Geology and plant life: the effects of landforms and rock types on plants. Seattle, University of Washington Press.

Martins, F.Q. & Batalha, M.A. 2006. Pollination systems and floral traits in cerrado woody species of the Upper Taquari region (central Brazil). Brazilian Journal of Biology 66: 543-552.

Meguro, M.; Pirani, J.R.; Giulietti, A.M. & Mello-Silva, R. 1994. Phytophysiognomy and composition of the vegetation of Serra do Ambrósio, Minas Gerais, Brazil. Revista Brasileira de Botânica 17: 149-166.

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.

Mera, G.; Hagen, M.A. & Orellana, J.A.V. 1999. Aerophyte, a new life form in Raunkiaer's classifi cation? Journal of Vegetation Science 10: 65-68.

Mourão, A. & Stehmann, J.R. 2007. Levantamento da flora do campo rupestre sobre canga hematítica couraçada remanescente na mina do Brucutu, Barão de Cocais, Minas Gerais. Rodriguésia 58: 775-786.

Mueller-Dombois, D. & Ellenberg, H. 1974. Aims and methods of vegetation ecology. New York, John Wiley & Sons.

Nimer, E. & Brandão, A.M.P.M. 1989. Balanço hídrico e clima da região dos Cerrados. Rio de Janeiro, Instituto Brasileiro de Geografia e Estatística.

Pirani, J.R.; Mello-Silva, R. & Giulietti, A.M. 2003. Flora de Grão-Mogol, Minas Gerais, Brasil. Boletim de Botânica da Universidade de São Paulo 21: 1-24.

Porembski, S. & Barthlott, W. 2000. Inselbergs - biotic diversity of isolated rock outcrops in tropical and temperate regions. Berlin, Springer Verlag.

Ramírez, N. 2004. Ecology of pollination in a tropical Venezuelan savanna. Plant Ecology 173: 171-189.

Ribeiro, K.T. & Medina, B.O. 2002. Estrutura, dinâmica e biogeografi a das ilhas de vegetação sobre rocha do Planalto do Itatiaia, RJ. Boletim do Parque Nacional do Itatiaia 10: 1-84.

Rizzini, C.T. 1997. Tratado de fitogeografia do Brasil: aspectos ecológicos, sociológicos e florísticos. 2 ed. Rio de Janeiro, Âmbito Cultural Edições.

Safford, H.D. & Martinelli, G. 2000. Southeast Brazil. Pp. 339-389. In: Porembski, S. & Barthlott, W. (Eds.). Inselbergs - biotic diversity of isolated rock outcrops in tropical and temperate regions. Berlin, Springer Verlag.

Sarmiento, G. & Monasterio, M. 1983. Life forms and phenology. Pp. 79-108. In: Bourlière, F. (Ed.). Ecosystems of the world: tropical savannas. Amsterdam, Elsevier.

Simmons, G.C. 1963. Canga caves in the Quadrilátero Ferrífero, Minas Gerais. Brazilian National Speleological Society Bulletin 25: 66-72.

Szarzynski, J. 2000. Xeric islands: environmental conditions on inselbergs. Pp. 37-48. In: Porembski, S. & Barthlott, W. (Eds.). Inselbergs - biotic diversity of isolated rock outcrops in tropical and temperate regions. Berlin, Springer Verlag.

van der Pijl, L. 1982. Principles of dispersal in higher plants. New York, Springer Verlag.

Viana, P.L. 2008. A flora dos campos rupestres sobre canga no Quadrilátero Ferrífero. Pp. 15-29. In: Simpósio sobre afloramentos ferruginosos no Quadrilátero Ferrífero: biodiversidade, conservação e perspectivas de sustentabilidade. Belo Horizonte, Anais.

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.

Vincent, R.C. & Meguro, M. 2008. Influence of soil properties on the abundance of plant species in ferruginous rocky soils vegetation, southeastern Brazil. Revista Brasileira de Botânica 31: 377-388.

Waser, N.M.; Chittka, L., Price, M.V.; Williams, N.M.; Ollerton, J. 1996. Generalization in pollination systems, and why it matters. Ecology 77: 1043-1060.

Recebido em 14/10/2010

Aceito em 29/04/2011

Appendix

Publication Dates

• Publication in this collection
  14 Sept 2011
• Date of issue
  June 2011

History

• Received
  14 Oct 2010
• Accepted
  29 Apr 2011