Flora 207 (2012) 662–672 Contents lists available at SciVerse ScienceDirect Flora journal homepage: www.elsevier.de/flora Floristics and environmental factors determining the geographic distribution of epiphytic bromeliads in the Brazilian Atlantic Rain Forest Talita Fontoura a,∗ , Veridiana Vizzoni Scudeller b , Andrea Ferreira da Costa c a Universidade Estadual de Santa Cruz, Departamento de Ciências Biológicas, Rodovia Ilhéus-Itabuna km16, 45662-900 Ilhéus, BA, Brazil Universidade Federal do Amazonas, Instituto de Biociências – ICB, Departamento de Biologia – Laboratório de Botânica, Av. General Rodrigo Octávio Jordão Ramos, 3000, Campus Universitário, Coroado I, Manaus, AM, Brazil c Museu Nacional/UFRJ, Departamento de Botânica, Quinta da Boa Vista, São Cristóvão, 20940-040 Rio de Janeiro, RJ, Brazil b a r t i c l e i n f o Article history: Received 25 September 2011 Accepted 1 May 2012 Keywords: Bromeliaceae Endemism Diversity Biogeography a b s t r a c t We investigated how epiphytic species and subfamilies of Bromeliaceae change along the extent of the Atlantic Rain Forest, to answer the questions: (i) How do the epiphytic genera and subfamilies of Bromeliaceae change along the domain? (ii) How similar are the different regions of the Atlantic Rain Forest in relation to the epiphytic species of bromeliads? (iii) Which environmental variables are the most important factors in determining species composition along the domain? We found 114 species of Bromelioideae and 73 of Tillandsioideae. The predominance of Bromelioideae was unexpected, because they are not wind-dispersed as would be expected for most epiphytes. The smaller number of species of Tillandsioideae, and the high frequency of species of Vriesea with limited geographic distributions indicated that epiphytes with rather limited geographic distributions predominate in this domain. Species similarity was divided into one block of south–southeastern localities, and a second block of northeastern–southeastern localities. These results suggest that the distribution of epiphytic bromeliad species resembles that of the phorophyte trees, more than a previous pattern suggested for all epiphytes in the domain. Latitude, temperature and altitude were important factors affecting the species composition along the domain. In general, our results differ from those of other studies in Latin America, and we suggest that historical and evolutionary events generated these differences. © 2012 Elsevier GmbH. All rights reserved. Introduction General distributional patterns of epiphytes were described in the seminal work of Gentry and Dodson (1987). Based on the number of epiphytic species and individuals found at different sites in Ecuador, these authors emphasized that one of the most striking distributional patterns shown by these plants is a tremendous decrease in the numbers of both species and individuals in drier habitats (Gentry and Dodson, 1987). However, an increase in humidity does not translate into a linear response in the number of epiphytic species. Rather, epiphytes are best developed at midelevations (Ibisch et al., 2001; Kessler, 2000, 2001, 2002a,b; Krömer and Kessler, 2005; Wolf and Flamenco-S, 2003), where moisture supply is usually continuous (Benzing, 1990). Several other investigations have explored large-scale or regional aspects of the distribution of these plants in northern and ∗ Corresponding author. Tel.: +55 73 3680 5105; fax: +55 73 3680 5226. E-mail addresses: fontoura.talita@gmail.com (T. Fontoura), scudellerveridiana@hotmail.com (V.V. Scudeller), afcosta@acd.ufrj.br (A.F.d. Costa). 0367-2530/$ – see front matter © 2012 Elsevier GmbH. All rights reserved. http://dx.doi.org/10.1016/j.flora.2012.05.003 western regions of South America (Ibisch et al., 2001; Kessler, 2000, 2001, 2002a,b; Krömer and Kessler, 2005; Küpper et al., 2004). For the eastern coast of Brazil the floristics and biogeography of epiphytes have been investigated on a large scale only recently (Kersten, 2010; Menini-Neto et al., 2009a). According to Kersten (2010) the Atlantic Rain Forest harbors approximately 2250 epiphytic species. The analysis of Menini-Neto et al. (2009a) focused on the floristic similarities of epiphytes sensu lato (see Benzing, 1990) in southern and south-eastern Brazil. According to these authors, the greatest similarity amongst the analyzed species distributions occurs in two groups. The first includes species of south-eastern Brazil, and a second group is composed by species of southern Brazil. However, investigations on the arboreal flora indicated a different biogeographic pattern than that described for epiphytes when areas from the north-eastern region of the Atlantic Rain Forest were taken into account. For arboreal plants, there is a floristic continuum between forests in southern Bahia (north-eastern Brazil) and the state of Rio de Janeiro in south-eastern Brazil (OliveiraFilho et al., 2005). Thus, an analysis of the epiphytic flora taking into account also north-eastern localities could indicate whether the floristic situation of epiphytes resembles the patterns that are T. Fontoura et al. / Flora 207 (2012) 662–672 seen in arboreal plants. In addition, further analysis focusing on species with a mainly epiphytic habit could clarify the extent to which the proposed biogeographic pattern of Menini-Neto et al. (2009a) is valid for species that depend solely on accumulated debris and throughfall, for instance the epiphytes of the family Bromeliaceae. The Bromeliaceae is the second largest family in terms of epiphytic species (Gentry and Dodson, 1987), and contains some 2500 species with this habit (Benzing, 1990). Although epiphytic bromeliads are characteristic of many humid areas of the Brazilian Atlantic Rain Forest (see Rizzini, 1992), much of the knowledge about these plants in this domain remains embedded in general studies of the family (Costa and Wendt, 2007; Fontoura et al., 1991; Leme and Siqueira-Filho, 2006a,b; Pontes and Agra, 2006; Sousa and Wanderley, 2000; Versieux and Wendt, 2006, 2007; Wanderley and Martins, 2007), or is based on local studies of epiphytes (see Alves et al., 2008; Menini-Neto et al., 2009a for references). These investigations are heavily biased toward the southern region of the Atlantic Rain Forest, where studies began long ago (Hertel, 1949). In the Bromeliaceae, epiphytic species are concentrated in the subfamilies Tillandsioideae and Bromelioideae (Smith and Downs, 1974, but see also Kessler, 2002a) which have distinct morphologies and dispersal modes. Tillandsioideae produce dehiscent capsules and the seeds are wind-dispersed, whereas the Bromelioideae produce berry-type fruits with animal-dispersed seeds (Smith and Downs, 1974). Local surveys in the Atlantic Rain Forest domain (e.g., Amorim et al., 2008, 2009; Fischer and Araújo, 1995; Kersten and Silva, 2001; Waechter, 1998) indicate that species of Tillandsioideae predominate in the southern region of Brazil, and that members of Bromelioideae predominate in the northern region. However, based on the different environmental factors affecting the floristics of the Atlantic Rain Forest (Marques et al., 2010; Oliveira-Filho and Fontes, 2000; Rizzini, 1992; Stehman et al., 2009), it seems unreasonable to expect that the subfamilies and species of epiphytes change based on only one environmental gradient. For instance, a floristic analysis of 125 localities in the domain indicated that the distance from the ocean is correlated with seasonality, clearly separating the effect of rainfall from that of semideciduous forest environments (Oliveira-Filho and Fontes, 2000). In addition, the floras are influenced by altitude, temperature range, annual mean temperature, and rainfall distribution (Oliveira-Filho and Fontes, 2000). Thus, the analysis of the epiphytes in southern, south-eastern and north-eastern regions of this domain in relation to certain environmental variables may help us to understand the flora and how it is affected by environmental variables in different areas of this domain. Additionally, we compare our results with other Latin American areas for which biogeographic studies exist. We characterized epiphytic bromeliad assemblages along the domain, in order to answer the following questions: (i) How do the epiphytic genera and subfamilies of Bromeliaceae change along the domain? (ii) How similar are the different areas of the Atlantic Rain Forest in relation to the epiphytic species of bromeliads? (iii) Which environmental variables are the most important factors in determining species composition along the domain? Methods Data matrix preparation We used 25 surveys performed in the Atlantic Rain Forest to analyze epiphytic bromeliads along the domain (Table 1). Nine surveys represented the southern region, 11 the south-eastern, and 5 the 663 north-eastern region of the domain (Table 1). For the Una region, two surveys (Amorim et al., 2008; Fontoura and Santos, 2010) were combined into one in the data matrix because both inventories were carried out in the same region but in different periods of time, using different field methodologies. Because of the large number of synonymies and new species published for the state of Bahia, all species names and identifications were verified by examination of voucher specimens in the following herbaria: Centro de Estudos e Pesquisa do Cacau (CEPEC), Herbarium Bradeanum (HB), Instituto de Botânica (IB), Jardim Botânico do Rio de Janeiro (RB), Museu Nacional do Rio de Janeiro (R), and Universidade Estadual de Santa Cruz (HUESC). Because Aechmea lingulata (L.) Baker has been split into 16 species (Leme and Siqueira-Filho, 2006a), all records of this species in the surveys analyzed were excluded, in order to avoid imprecision in the results due to voucher misidentification. This decision was taken in order to avoid false similarities in the multivariate analysis. Even though specificity of relationships between epiphyte species and epiphyte phorophytes is rarely recorded (Benzing, 1990; Laube and Zotz, 2006), we avoided epiphyte surveys based only on a small number of phorophytic species (Gonçalves and Waechter, 2002, 2003; Reis and Fontoura, 2009), in order to maximize species richness in the sampled areas. General aspects We characterized epiphytic bromeliads along the domain based on the percentage of subfamilies in each locality and in each geopolitical region. We also used the richest genus in each locality to evaluate general patterns. Because the inventories used different areas, methodologies and field efforts, it was expected that only the most major changes in subfamilies and genera could be detected along the domain. Data analysis Species similarities were analyzed using cluster and ordination analyses based on a binary matrix of 184 species × 25 localities (Appendix 1). Because 102 species were restricted to only one locality, a second matrix of 82 species × 25 localities was used in the multivariate analysis, to assess possible modifications in the results because of single species. The program FITOPAC 1 (Shepherd, 1995) was used to calculate the Jaccard coefficient amongst sample units (localities) and to generate a dendrogram using the unweighted pair group method using arithmetic averages (UPGMA; Sneath and Sokal, 1973). The following environmental variables were used as possible predictors of the epiphyte community along the domain: (i) mean annual rainfall, (ii) distance from the ocean, (iii) altitude, (iv) latitude, (v) mean temperature, (vi) minimum mean temperature, and (vii) relative humidity. Rainfall, altitude and mean temperature were based on the information available in the epiphyte surveys. The variable “distance from the ocean” was obtained by plotting all localities on a map using the Geographic Information System ArcGis ver. 9.2. Minimum mean temperature and relative humidity were obtained from www.leb.esalq.usp.br/angelocci/. The main matrix (184 species × 25 localities) and the secondary matrix (25 localities × 7 environmental variables) were analyzed using Canonical Correspondence Analysis (CCA), running the PCOrd program ver. 3.11 (McCune and Mefford, 1997). The Monte Carlo test was performed using 1000 iterations, to assess the significance values of correlations. Biplot graphs were interpreted following Ter Braak (1997). Surveys (codes of survey areas; for locations see Fig. 1) Geopolitical region Area (ha) Sampling method Vegetation type and reference BA-1 BA-2 BA-3 BA-4 ES MG-1 MG-2 NE NE NE NE SE SE SE 13(57.1) 29(59.2) 19(55.9) 14(60.9) 21(67.7) 6(46.2) 7(31.8) 9(42.9) 20(40.8) 15(44.1) 9(39.1) 10(32.3) 7(53.8) 15(68.2) 21 49 34 23 31 13 21 300 21 500 6000 7500 0.34 n.i. n.i. Random Random Random Random 34 samples of 10 m × 10 m Random Random 8(80) 22(56.4) 5(45.5) 10 39 11 0.9 7500 1250 0 2(50) 11(40.7) 7(31.8) 8(47.1) 10(34.4) 1(6.3) 2(22.2) 3(33.3) 3(100) 2(50) 16(59.3) 15(68.2) 9(52.9) 21(65.6) 15(93.8) 7(77.8) 6(66.6) 3 4 27 22 17 32 16 9 9 22 500 22 500 22 500 Random Random 119 phorophytes sampled in point center quarter method 320 m × 320 m sample plot 320 m × 320 m sample plot 320 m × 320 m sample plot Seven transects: 600–2400 m2 Transect 3 m × 2 km 320 m × 320 m sample plot 30 samples of 100 m2 110 random phorophytes Random and 90 phorophytes Highland ombrophilous dense foresta Lowland ombrophilous dense forestb , c Highland ombrophilous dense foresta Highland ombrophilous dense foresta Highland ombrophilous dense forestd Seasonal semideciduous montane and gallery forestse Gallery forest, evergreen upper montane forest, and seasonal montane forest grovese Alluvial ombrophilous dense forestf Highland ombrophilous dense forestg Restinga foresth MG-3 RJ-1 RJ-2 SE SE SE 2(20) 17(43.6) 6(54.5) SP-1 SP-2 SP-3 SP-4 SP-5 SP-6 PR-1 PR-2 PR-3 SE SE SE SE SE SE S S S SC-1 S 4(33.3) 8(66.7) 12 n.i. RS-1 RS-2 RS-3 RS-4 S S S S 4(40) 2(33.3) 2(16.7) 2(16.7) 6(60) 4(66.7) 10(83.3) 10(83.3) 10 6 12 12 n.i. n.i. 20 180 1 RS-5 S 2(22.2) 7(77.8) 9 Species number of subfamilies (%) Bro c d e f g h i j k l m n o p q r s t Amorim et al. (2009). Amorim et al. (2008). Fontoura and Santos (2010). Varassin (2002). Menini-Neto et al. (2009a). Menini-Neto et al. (2009b). Fontoura et al. (1997). Fontoura and Reinert (2009). Breier (2005). Santos (2000). Fischer and Araújo (1995). Kersten and Silva (2001). Kersten and Silva (2002). Kersten et al. (2009). Bonnett and Queiroz (2006). Rogalski and Zanin (2003). Buzatto et al. (2008). Oliveira et al. (2005). Waechter (1998). Giongo and Waechter (2004). Til 0.6 22 500 0.3 8.6 n.i. n.i. 240 phorophytes sampled in point center quarter method Random Random Random 60 phorophytes in point center quarter method 60 phorophytes sampled in point center quarter method Seasonal semideciduous foresti Savannah vegetationi Ombrophilous dense foresti Dense forestj Riparian, dense and restinga forestk Restinga foresti Ombrophilous dense forest and restingal Mixed ombrophilous alluvial forestm Mixed ombrophilous alluvial and high-altitude montane forestn Ombrophilous dense foresto Semideciduous forestp Mixed ombrophilous forestq Semideciduous seasonal forestr Restinga forests Riparian forestt T. Fontoura et al. / Flora 207 (2012) 662–672 a b Total number of Bromeliaceae 664 Table 1 General aspects of field surveys along the Brazilian Atlantic Rain Forest. Geopolitical regions: NE – northeastern, SE – southeastern, S – south. State codes: BA – Bahia, ES – Espírito Santo, MG – Minas Gerais, RJ – Rio de Janeiro, SP – São Paulo, PR – Paraná, SC – Santa Catarina, RS – Rio Grande do Sul. Subfamilies of Bromeliaceae: Bro – Bromelioideae, Til – Tillandsioideae. n.i., not indicated in reference. T. Fontoura et al. / Flora 207 (2012) 662–672 665 Table 2 Surveys (codes), total amount of epiphytic Bromeliaceae species (richness), and richness of the richest genus occuring in the Brazilian Atlantic Rain Forest. Survey 1. BA-1 2. BA-2 3. BA-3 4. BA-4 5. MG-1 6. MG-2 7. MG-3 8. ES 9. RJ-1 10. RJ-2 11. SP-1 12. SP-2 13. SP-3 14. SP-4 15. SP-5 16. SP-6 17. PR-1 18. PR-2 19. PR-3 20. SC 21. RS-1 22. RS-2 23. RS-3 24. RS-4 25. RS-5 Richness 21 49 34 23 13 21 10 31 39 11 3 4 27 22 17 32 16 9 9 12 10 6 12 12 9 Genera (number of species) Vriesea (6) Vriesea (11) Vriesea (10) Vriesea (7) Vriesea (4) Vriesea (7) Tillandsia (4), Vriesea (4) Aechmea (9) Vriesea (17) Aechmea (3), Tillandsia (3) Tillandsia (3) Tillandsia (2) Vriesea (12) Vriesea (10) Vriesea (8) Vriesea (14) Vriesea (11) Tillandsia (6) Tillandsia (4) Tillandsia (4), Vriesea (4) Tillandsia (5) Tillandsia (4) Tillandsia (6) Tillandsia (5), Vriesea (5) Tillandsia (6) Floristic similarities amongst localities Fig. 1. Location of the surveys along the Brazilian Atlantic Rain Forest; dominant subfamilies and richest genus: asterisks – higher percentage of Tillandsioideae species, dots – higher percentage of Bromelioideae; A – Aechmea, T – Tillandsia, V – Vriesea. Results General aspects The 184 species analyzed comprised 112 Bromelioideae and 72 Tillandsioideae, i.e., 1.6 times more Bromelioideae than Tillandsioideae. In general, the number of species decreased toward the southern region of the domain (34 species) where Tillandsioideae reached the highest percentage (76.5%). The predominance of this subfamily remains high from the southern region to the inland (MG-1, 2, 3) and the mountainous areas (RJ-1) of the south-eastern region of the domain (Fig. 1, Table 1). Bromelioideae predominated from the restinga forest of the south-eastern region (RJ-2) northwards to all other localities (Fig. 1). The north-eastern (83 species) and south-eastern (117 species) regions harbored the highest percentages of Bromelioideae (63.9% and 53.8%, respectively). Vriesea was the most frequent genus, occurring in 15 (60%) localities, followed by Tillandsia, which occurred in nine (36%, Table 2). Although species of Tillandsia occur in some localities of the south-eastern region, the genus is mainly represented in the southern-most region and at inland localities of the Atlantic Rain Forest. Vriesea predominated from the northeast until parts of the southern region, and in localities near the ocean (Fig. 1). The cluster analysis resulted in two large groups and five subgroups with similarity values around 0.1 (Fig. 2a). The first group was composed by three subgroups with localities in the southern and south-eastern regions of the Atlantic Rain Forest (Fig. 2a). The localities of the first subgroup were characterized by the presence of Tillandsia recurvata, T. stricta, and T. usneoides in at least nine of 12 localities. The second subgroup comprised seven localities, where Vriesea carinata, V. flammea, and V. philippocoburgii were shared species in six or all localities of this subgroup. The third subgroup consisted only of SP-2, where B. zebrina and T. recurvata were shared with seven and nine other localities of this first, large group. The second group was composed by two subgroups, with one locality from the south-eastern and all the surveys from the northeastern region (Fig. 2a). All the surveys in southern Bahia formed the fourth subgroup. Twelve species occurred in at least three localities of this subgroup: B. saundersii, Neoregelia kerryi, Nidularium bicolor, Ni. innocentii, Racinaea spiculosa, T. meridionalis, V. vagans, V. ensiformis, V. psittacina, V. flammea, V. minuta, and V. simplex. The fifth subgroup consisted of the Espírito Santo survey, with seven species in common with three or four other localities of this group: Aechmea nudicaulis, Portea petropolitana, R. spiculosa, T. tenuifolia, V. drepanocarpa, V. psittacina, and V. simplex. The cluster analysis with 82 species, excluding singletons, formed the same two large groups and subgroups as shown in the full analysis (Fig. 2b). The only differences occurred in the linkage level of the subgroups. Environmental variables The high eigenvalues (>0.5) on CCA axes 1, 2, and 3 indicate that much of the variation in the species data is explained by these axes (Table 3), and considerable turnover of species occurs along these axes. The cumulative variance of 21.6% on the third axis indicates that considerable noise remained unexplained by the CCA. Correlation values between species and environmental variables were consistently high (all >0.96) and significant values were obtained on the first and second axis (Table 3), indicating 666 T. Fontoura et al. / Flora 207 (2012) 662–672 Fig. 2. Cluster analysis of epiphytic bromeliad species along the Brazilian Atlantic Rain Forest domain. Localities divided in first group (G1) and second group (G2). South surveys denoted in gray, southeastern in black, and northeastern surveys in italics. (A) Dendrogram based on 184 species and 25 localities along the domain. (B) Dendrogram based on 82 species and 25 localities. Table 3 Summary statistics of CCA and P values of Monte Carlo tests. Data matrix of 187 bromeliad epiphytic species × 25 localities along the Brazilian Atlantic Rain Forest. Eigenvalue Cumulative variance explained in species data Pearson correlation between species and environmental variables P values for eigenvalues P values for correlations between species and environmental variables Axis 1 Axis 2 Axis 3 0.711 9.1 0.992 0.001 0.002 0.524 15.8 0.977 0.006 0.01 0.459 21.6 0.970 0.014 0.09 T. Fontoura et al. / Flora 207 (2012) 662–672 Table 4 Intraset correlations between environmental and geographical variables and CCA axis. Data matrix of 187 bromeliad epiphytic species × 25 localities along the Brazilian Atlantic Rain Forest. Highest values denoted in bold. Variable Axis 1 Axis 2 Axis 3 Latitude Altitude Distance from ocean Rainfall Mean temperature Minimum temperature Relative humidity 0.945 0.078 0.385 0.061 −0.792 −0.875 −0.413 0.063 −0.543 −0.483 0.438 0.199 −0.176 0.341 −0.138 0.685 −0.150 0.349 −0.266 −0.228 0.069 that results can be interpreted based on the variation of environmental variables of both axes of the biplot. The high values of correlations (r > 0.6) between environmental variables and the CCA axes (Table 4) indicate that latitude and mean and minimum temperatures are important gradients on axis 1. The most important gradient on axis 3 of the biplot is represented by altitude. 667 All localities of Bahia, Espírito Santo, and RJ-2 were segregated in the left region of the biplot, and all other localities were positioned in the right region (Fig. 3). The projection of localities onto the vectors indicated the degree to which these variables influenced the species composition of epiphytes in the different localities of the Atlantic Rain Forest. Latitude moderately influenced the species composition of MG-1, RJ-1, SP-3, 5, 6, PR-1, and influenced species composition of SP-4, PR-2, SC-1 and RS-4. The distance from the ocean and, in lesser degree, the altitude influenced epiphytic community of localities MG-2, 3, SP-1, 2, PR-3 and RS-1, 2, 3, 4, 5. Temperature was an important factor for epiphytes in all localities of Bahia and Espírito Santo, and moderately influenced epiphytes of RJ-2 (Fig. 3). Discussion General aspects Our analysis added further details on how members of the Bromelioideae predominate in the Atlantic Rain Forest, and how Axis 2 SP-6 1,0 SP-3 SP-5 SP-4 PR-1 BA-2 PR-2 RH RJ-1 Tme MG-1 Lat BA-4 Tmi -1,5 BA-1 -0,5 BA-3 0,5 Axis 1 RS-4 SC-1 1,5 RJ-2 Alt RS-5 DistOce RS-3 PR-3 RS-2 MG-2 -1,0 MG-3 RS-1 SP-1 ES SP-2 -2,0 Fig. 3. Biplot of Canonical Correspondence Analysis based on 184 species of epiphytic bromeliads in 25 localities along the Brazilian Atlantic Rain Forest. Species names were omitted to preserve clarity. Rainfall eigenvector is clear only to the third axis (not shown). Environmental variables: Lat – latitude, DistOce – distance from the ocean, Alt – altitude, Tmi – mean of minimum temperature, Tme – mean temperature, RH – relative humidity. 668 T. Fontoura et al. / Flora 207 (2012) 662–672 the main subfamilies of epiphytic species shift along the domain. In the surveys, Bromelioideae comprised 1.6 times more epiphytic species than Tillandsioideae. For the family as a whole (see Martinelli et al., 2008), Bromelioideae comprises 2.3 times more species than Tillandsioideae. The preponderance of this subfamily in both studies indicates that the Bromelioideae generally surpasses the Tillandsioideae in number of species, considering either all life forms or only the epiphytic species. If our findings represent the general pattern of subfamilies of epiphytic bromeliads in the domain, this implies that most members of Bromeliaceae in the Atlantic Forest should be animal-dispersed epiphytes, according to the fruit type. This pattern contrasts with bromeliads of the Bolivian Andes, which are mainly wind-dispersed species (Kessler, 2002a). In the Brazilian Atlantic Rain Forest, the highest percentage of Tillandsioideae is located in the southern region, where the overall species richness is lower. The variety of sampling methods and areas amongst the different surveys prevents detailed comparisons of raw richness numbers. However, it is noteworthy that four field surveys in the north-eastern region collected 2.4 times as many species as did nine surveys in southern Brazil, and 0.7 times fewer species than 12 surveys in south-eastern Brazil. Even though our data suggest that differences in field efforts did not affect the general pattern of low richness in the southern region, field surveys using the same methodology along the domain would be useful to measure more precisely modifications in alpha, beta and gamma diversity along different gradients (see Cody, 1993). The relatively few species in southern Brazil were mainly members of Tillandsioideae, which suggests that the low winter temperatures in this region may limit species richness. In addition, if we consider that Tillandsioideae evolved in northern South America (Barfuss et al., 2005; Givnish et al., 2007), then the geographic distance between the center of diversity of this subfamily and southern Brazil may be a second factor contributing to the rather low richness of Tillandsioideae in this region. The prominence of Tillandsioideae in the south is very clear until the mountainous region of Rio de Janeiro state. However, from the restinga forest, which at its closest point is located approximately 53 km from the mountains, to the more northern regions of the domain, there is a shift in the predominant subfamily. This turnover of species suggests that the region of Rio de Janeiro may have an important biogeographic role, where both subfamilies co-occur in the different habitats of this state. The occurrence of higher percentages of Tillandsioideae in most localities of the Atlantic Rain Forest indicates that the mesic Vriesea and the Tillandsia species which depend on atmospheric humidity, exploit different habitats establishing and increasing their species number along the domain. Regardless of the different methodologies and field efforts used in the inventories, it is clear that mountainous areas of Serra do Mar (represented by RJ-1, SP-3, 4, 5, 6) and several other regions (e.g., MG-1, 2) are dominated by Vriesea species. Thus, if the bromelioid epiphytic clade in and around the Serra do Mar originated roughly 5.5 Ma (Givnish et al., 2011), it would be interesting to investigate which group of Bromeliaceae first arrived in this mountainous region. Nowadays, members of Tillandsioideae are overwhelmingly represented. Although Gentry and Dodson (1987) and Wolf and Flamenco-S (2003) stated that wind-dispersed epiphytes are better represented in relatively dry forests, this pattern does not apply well to the Atlantic Rain Forest. Wind-dispersed epiphytic bromeliads dominate in different parts of the south-eastern and southern regions of the domain, either in humid (e.g., SP-5, 6) or in drier locations of the domain (SP-2, RS-1). Actually, our data indicate that Tillandsia is the richest genus in the localities farthest inland, and Vriesea is the richest genus in localities near the Atlantic Ocean. Although several investigations have explored how Bromelioideae have evolved and gained their rosulate and epiphytic habit (Sass and Specht, 2010; Schulte et al., 2005, 2009), it is virtually unknown how members of Tillandsia and Vriesea succeeded in occupying such a large area of humid forest in the Brazilian Atlantic Rain Forest. For the other parts of the domain, our data indicate that animals have played a prominent role in dispersing epiphytic bromeliads. This is an unexpected result in terms of epiphytism, because usually wind is the most common dispersal mode amongst epiphytes (Gentry and Dodson, 1987). The dominance of Bromelioideae begins in the restinga forest of Rio de Janeiro (RJ-2), increases toward the drier (see also Amorim et al., 2005; Reis and Fontoura, 2009) and humid areas of the northeastern region (e.g., BA-2), and decreases toward the south-eastern and southern regions of Brazil. The various ecosystems, rainfall amounts, and altitudes indicate that the dominance of this subfamily is not related to one or another habitat, but rather to evolutionary trends and to the means by which the domain was occupied by this life form. Apparently, the southern-most regions of the domain are unfavorable for the establishment of many bromelioid epiphytes. As the members of this subfamily are exclusively animal-dispersed, possibly the available vectors were or still are unable to disperse epiphytic bromelioids along the domain. Perhaps the refuge theory (Whitmore and Prance, 1987) may explain all or part of the observed pattern. However, further studies based on palaeopalynological and phylogenetic investigations may reveal to what degree this theory is valid for this region with high species richness and endemism. Our data for the most species-rich genera in the domain, added to the data contained in Martinelli et al. (2008), Smith and Downs (1974), and Stehmann et al. (2009), suggest intriguing questions about the geographic range of epiphytism. Vriesea is the largest genus in the Atlantic Rain Forest (Martinelli et al., 2008), and about 87% of the Vriesea species in this domain are epiphytic (see Smith and Downs, 1977; Stehmann et al., 2009). Added to these facts, our data show that Vriesea is also the most frequent genus occurring in almost all Brazilian states. Of the 166 Vriesea species in the domain, only 3.5% occur throughout its extent, and 70% occur in only one or two Brazilian states (see Martinelli et al., 2008); i.e., the great majority of species have a rather restricted geographic distribution. Therefore, even if surveys were regularly distributed in forested areas to allow the collection of this highly epiphytic genus, the geographic distribution of species would remain biased toward a narrow geographic distribution, reflecting the distributions of Vriesea species. Although quantitative investigations are highly desirable, the evaluation of the geographic distributions of our 187 analyzed species, added to the data provided by Martinelli et al. (2008) and Versieux and Wendt (2007), reinforces the hypothesis that epiphytic bromeliads of this domain probably have generally smaller ranges than those occurring in the Bolivian Andes (Kessler, 2002a). Seventy-six species (40.6%) occur in only one or two Brazilian states (see Martinelli et al., 2008), mainly represented by localities in southern Bahia and Espírito Santo, representing a geographic distribution of approximately 14 squared degrees. Twenty-four species (12.8%) occur in Minas Gerais, Espírito Santo and Rio de Janeiro. If these 24 species would occupy this geographic range completely, they would cover in total an area of 74 squared degrees (see Versieux and Wendt, 2007). These figures are different from those shown by Kessler (2002a), which indicated a mean range size of 147 squared degrees for the epiphytic bromeliads of the Bolivian Andes. Further quantitative investigations may confirm whether bromeliads of the Atlantic Rain Forest have as extensive a geographic distributions as those reported by Kessler (2000, 2001, 2002a,b), and by Givnish et al. (2007) for epiphytes in general. T. Fontoura et al. / Flora 207 (2012) 662–672 Floristic similarities Both multivariate analyses resulted in very similar clusters, indicating that the exclusive species had no effect on the cluster analysis. The analyses also revealed that the similarities of groups and subgroups of epiphytic bromeliads along the domain were not based on the dichotomy between the south-eastern and southern regions, as observed for the epiphytes as a whole (Menini-Neto et al., 2009a). With the available data for epiphytic plants, it also cannot be said that the north-eastern region is strongly dissimilar to other regions of this domain, as occurs for arboreal plants (Scudeller et al., 2001). The number of shared species amongst localities was apparently the determining factor in the division of groups and subgroups. This characteristic is especially clear when the two main groups are compared. In the first group, all subgroups have three to four shared species amongst many localities in the southern and part of the south-eastern region of the domain. In contrast, the second group has seven to 12 shared species amongst localities in Espírito Santo and southern Bahia. Actually, the resulting dendrogram and main divisions of localities indicate that the general pattern of epiphytic bromeliads along the domain conforms to the pattern known as the “Campos dos Goytacazes gap” as described by Oliveira-Filho and Fontes (2000) for arboreal plants. According to this pattern, the low rainfall in northern Rio de Janeiro state has led to the development of a semi-deciduous forest. Apparently, in their history the epiphytic bromeliads were sensitively affected by the low rainfall of this region, and the net result of such effects is reflected in our dendrograms that clearly separate the species in southern Bahia and Espírito Santo from all other localities of the domain. Although additional field surveys are necessary to better compare the floras of Rio de Janeiro and Espírito Santo, it is interesting to note that the epiphyte flora of mountainous areas of Espírito Santo is dissimilar to the flora of the Rio de Janeiro mountains – which is different from the situation of arboreal plants (OliveiraFilho et al., 2005). Most probably, it was indeed the low rainfall in northern Rio de Janeiro that contributed to separate the epiphyte flora into two blocks. Environmental variables The CCA indicated that latitude is an important factor determining the species composition in almost all localities near the Atlantic Ocean, where rainfall occurs in all months of the year. In most of these localities, Vriesea is the most species-rich genus. The first exception is in extreme southern Brazil (RS-4), where Tillandsia and Vriesea share the highest richness in a restinga forest near the ocean. The second exception is represented by the mixed ombrophilous alluvial forest of PR-2 where Tillandsia is the most species-rich genus. Because the presence of Tillandsia is usually associated with high light intensities (Benzing, 1990, 2000), finer studies may reveal whether natural forest formation or human disturbances may have allowed the presence of this genus in both localities. However, it would be reasonable to state that latitude is an important factor where Vriesea is the richest genus – in other words, in most of the humid localities in south-eastern and southern Brazil. In these localities, other environmental variables have relatively little effect on the species composition. Apparently, species of this genus take advantage of the high levels of rainfall in these humid locations to accumulate water in their tanks, as a means to compensate for physiological dryness imposed by their occupation of tree crowns and trunks. Distance from the ocean and, in lesser degree, altitude were important variables affecting species composition in most localities where Tillandsia is the most species-rich genus. The first exception is represented by the surveyed areas in MG-2 (gallery forest, evergreen upper montane forest, and seasonal montane forest groves) 669 and the second, by the alluvial ombrophilous dense forest in MG3. It is probable that in spite of the large distance between these localities and the Atlantic Ocean, there is enough humidity in the gallery forest, evergreen upper forest, and in the alluvial forest to allow the establishment of several Vriesea species. In relation to the low contribution of altitude, it is probable that our large scale study encompassing almost 15◦ in latitude and 25◦ in longitude is too large to detect altitudinal patterns – that may be better analyzed at a smaller scale. The Brazilian Atlantic Rain Forest does not form a massif barrier of, e.g., mountain ridges (see Fiaschi and Pirani, 2009; Oliveira-Filho and Fontes, 2000) along the entire domain, and do not range from 0 to more than 3000 m.a.s.l., as is the altitudinal span of Bromeliaceae occurrences in the Andes (see Haffer, 1987; Küpper et al., 2004). In eastern Brazil the mountainous areas are scattered along the whole domain (Safford, 2007) and probably other local factors may be responsible for the species distribution, so that the altitude parameter has only low weight in our results. As noted by Oliveira-Filho and Fontes (2000), the influence of altitude is far more complex than simply creating temperature gradients and frost events. Rising elevation also decreases atmospheric pressure, increases solar radiation, accelerates air masses, promotes greater cloudiness, and increases rainfall (Jones, 1992). Although evaluation of other variables is necessary to better investigate the effects of environmental variables in drier locations, it is interesting to note that the occupation of these regions probably became possible because Tillandsia species became able in their evolution to use atmospheric humidity overcoming the dry periods characteristic of these more-inland regions of the Atlantic Rain Forest. Finally, in the correlations, minimum temperature turned out to be important affecting the species composition of localities at altitudes from 640 to 775 m.a.s.l. (BA-1, 3 and 4), and mean annual temperature resulted as an important factor for the lowland area of Una (BA-2), located about 100 m.a.s.l. For arboreal vegetation temperature can be one of the most important factors, shaping the world’s vegetation formations as well as the plant cover in southern Brazil (Oliveira-Filho and Fontes, 2000) where arboreal species richness is low. However, to epiphytic bromeliads both temperature variables are important especially in localities where, at low latitude, the majority of species are animal dispersed – probably in one of the richest areas of the domain with concern of the Bromeliaceae flora. Apparently, the association between epiphytic bromeliads and animals was evolutively profitable, promoting both species richness and endemism. Clearly, other variables ought to be taken into account, and for some parameters our scale of analysis, encompassing a large part of the Atlantic Rain Forest, perhaps may obscure processes that are better understood on smaller scales (e.g., pollination and dispersal processes). Moreover phylogeographic studies must complement now the statistical treatments of floristic relations in order to support or correct the biogeographic patterns presented here. Acknowledgments This paper was part of post-doctoral work of TF in the Departamento de Botânica do Museu Nacional/Universidade Federal do Rio de Janeiro. We are grateful to the curators of the Centro de Estudos e Pesquisa do Cacau (CEPEC), Herbarium Bradeanum (HB), Instituto de Botânica (IB), Jardim Botânico do Rio de Janeiro (RB), Museu Nacional do Rio de Janeiro (R), and Universidade Estadual de Santa Cruz (HUESC), and to the curators of the M and Gray Herbaria, who kindly provided images of Bromeliaceae from Bahia. We are also grateful for many suggestions from reviewers, which led to substantial improvements in the manuscript. T. Fontoura et al. / Flora 207 (2012) 662–672 670 Appendix 1. Species of epiphytic bromeliads of the Brazilian Atlantic Rain Forest used in the data matrix (184 species × 25 localities) as the basis of multivariate analyses Species Surveys (codes of survey areas; for locations see Fig. 1) Acanthostachys strobilacea (Schult. f.) Klotzsch Aechmea araneosa L.B. Sm. A. blanchetiana (Baker) L.B. Sm. A. bromeliifolia (Rudge) Baker A. bruggeri Leme A. caesia E. Morren ex Baker A. calyculata Baker A. capitata Baker A. cariocae L.B. Sm. A. castanea L.B. Sm. A. coelestis (K. Koch) E. Morren A. conifera L.B. Sm. A. depressa L.B. Sm. A. distichantha Lem. A. fasciata (Lindl.) Baker A. floribunda Mart. ex Schult. f. A. gamosepala Wittm. A. hostilis E. Pereira A. leonard-kentiana H.E. Luther & Leme A. leucolepis L.B. Sm. A. marauensis Leme A. miniata Baker A. mollis L.B. Sm. A. nudicaulis (L.) Griseb. SP-2, ES A. organensis Wawra A. ornata Baker A. pectinata Baker A. perforata L.B. Sm. A. pineliana (Brongn. ex Planch.) Baker A. ramosa Mart. ex Schult. f. A. recurvata (Klotzsch) L.B. Sm. A. sphaerocephala Baker A. tentaculifera Leme, Amorim & J.A. Siqueira A. triangularis L.B. Sm. A. turbinocalyx Mez A. victoriana L.B. Sm. A. viridostigma Leme & Luther Billbergia alfonsi-joannis Reitz B. amoena (Lodd.) Lindl. B. distachia (Vell.) Mez B. euphemiae E. Morren B. horrida Regel B. morelii Brongn. B. nutans H. Wendl. ex Regel B. pyramidalis (Sims) Lindl. B. sanderiana E. Morren B. saundersii Read B. tweedieana Baker B. vittata Brongn. ex Morel B. zebrina (Herb.) Lindl. Canistropsis billbergioides (Schult. f.) Leme Canistrum camacaense Martinelli & Leme C. montanum Leme C. seidelianum W. Weber C. triangulare L.B. Sm. & Reitz Catopsis berteroniana (Schult. & Schult. f.) Mez C. sessiliflora (Ruiz & Pav.) Mez Edmundoa lindenii (Regel) Leme Guzmania lingulata (L.) Mez Hohenbergia augusta (Vell.) E. Morren H. belemii L.B. Sm. & Read ES BA-2 ES MG-3 RJ-1 RS-1 BA-4, BA-3 BA-2 ES SP-3, RJ-1 BA-2 BA-2 PR-3, PR-2 RJ-2 RJ-2 SP-6, SP-5, SP-3 ES BA-2 BA-2 BA-2 BA-1 BA-2 SC-1, PR-1, SP-6, SP-5, SP-4, SP-3, ES, MG-2, BA-4, BA-3 SP-5, SP-4 SP-4, SP-3, RJ-1 SP-6, SP-5, SP-3 BA-2 RJ-1, ES MG-1, BA-1 RS-5, RS-4, RS-2, RS-1, PR-3, PR-2 RJ-2 BA-4, BA-3 ES BA-2, BA-1 ES BA-4, BA-3, BA-1 MG-2 SP-6, SP-4, SP-3, RJ-2, RJ-1, ES MG-3 MG-1, BA-3 BA-4 BA-2 RS-3, RS-2, RS-1, PR-3 RJ-2, RJ-1 RJ-1 BA-4, BA-3, BA-2, BA-1 MG-1 ES RS-5, RS-4, RS-3, RS-1, SC-1, SP-6, SP-2, MG-1 SP-3 BA-4, BA-3 BA-3, BA-1 BA-3, BA-1 ES PR-1 SP-6, BA-2 SC-1, SP-6, SP-4, SP-3, RJ-1 BA-3, BA-2 ES BA-4, BA-2 Species Surveys (codes of survey areas; for locations see Fig. 1) H. blanchetii (Baker) E. Morren ex Mez H. brachycephala L.B. Sm. H. edmundoi L.B. Sm. & Read H. hatschbachii Leme H. minor L.B. Sm. H. ramageana Mez Lymania alvimii (L.B. Sm. & Read) Read L. azurea Leme L. corallina (Beer) Read L. marantoides (L.B. Sm.) Read L. smithii Read Neoregelia azevedoi Leme N. carolinae (Beer) L.B. Sm. N. crispata Leme N. eltoniana W. Weber N. farinosa (Ule) L.B. Sm. N. fluminensis L.B. Sm. N. ibitipocensis (Leme) Leme N. kerryi Leme N. laevis (Mez) L.B. Sm. N. longisepala E. Pereira & Leme N. lymaniana R. Braga & Sucre N. magdalenae L.B. Sm. & Reitz N. oligantha L.B. Sm. N. pauciflora L.B. Sm. N. wilsoniana M.B. Foster Nidularium amorimii Leme N. antoineanum Wawra N. bicolor (E. Pereira) Leme N. cariacicaense (W. Weber) Leme N. ferdinandocoburgii Wawra N. innocentii Lem. BA-2 N. krisgreeniae Leme N. logiflorum Ule N. marigoi Leme N. microps E. Morren ex Mez N. procerum Lindm. N. rubens Mez N. rutilans E. Morren N. scheremetiewii Regel Portea filifera L.B. Sm. P. nana Leme & H.E. Luther P. petropolitana (Wawra) Mez Quesnelia arvensis (Vell.) Mez Q. clavata Amorim & Leme Q. humilis Mez Q. koltesii Amorim & Leme Q. lateralis Wawra Q. liboniana (De Jonghe) Mez Q. quesneliana (Brongn.) L.B. Sm. Q. strobilispica Wawra Q. testudo Lindm. Racinaea aeris-incola (Mez) M.A. Spencer & L.B. Sm. R. spiculosa (Griseb.) M.A. Spencer & L.B. Sm. Ronnbergia brasiliensis E. Pereira & L.A. Penna R. silvana Leme Tillandsia aeranthos (Loisel.) L.B. Sm. T. bulbosa Hook. T. crocata (E. Morren) Baker T. dura Baker T. gardneri Lindl. T. geminiflora Brongn. T. globosa Wawra T. kautskyi E. Pereira T. loliacea Mart. ex Schult. f. T. mallemontii Glaziou ex Mez T. meridionalis Baker T. pohliana Mez BA-2 BA-3 BA-2 BA-4 BA-2 BA-2 BA-2 BA-2 BA-1 BA-2 BA-3 RJ-1 BA-3, BA-2 RJ-2 MG-1 RJ-1 MG-2 BA-4, BA-3, BA-1 SP-6, SP-3 BA-2 MG-2 ES MG-2 ES, BA-3 BA-2 BA-2 SP-5 BA-4, BA-3, BA-1 ES MG-2 SC-1, SP-6, SP-5, SP-4, RJ-1, BA-4, BA-3, BA-2, BA-1 SP-3 RJ-1, ES, MG-1 MG-2 RJ-1 SP-6, RJ-1, ES, BA-2 SP-4 SP-3 RJ-1 BA-1 BA-3 ES, BA-4, BA-2 SP-6 BA-3 SP-5 BA-4 RJ-1 RJ-1 ES ES SP-5 MG-2, RJ-1 SP-6, SP-3, RJ-1, ES, BA-3, BA-2, BA-1 BA-3, BA-2 BA-1 RS-5, RS-4, RS-3, RS-2 BA-2 PR-2 SP-6 RS-4, RS-1, PR-1, SP-6, RJ-2, MG-3, MG-2, BA-2 RS-5, RS-4, RS-3, RS-1, SC-1, PR-1, SP-6, SP-4, RJ-1, ES, MG-2, MG-1 SP-6, SP-3 ES SP-1 SC-1, PR-2 BA-4, BA-3, BA-1 SP-2, BA-3 T. Fontoura et al. / Flora 207 (2012) 662–672 Species Surveys (codes of survey areas; for locations see Fig. 1) T. recurvata (L.) L. RS-5, RS-3, RS-2, RS-1, PR-3, PR-2, SP-2, SP-1, MG-3, MG-2 BA-2 RS-5, RS-4, RS-3, RS-2, RS-1, SC-1, PR-3, PR-2, SP-6, SP-5, SP-4, SP-3, SP-1, RJ-2, RJ-1, MG-2, MG-1, BA-4, BA-3, BA-2 RS-5, RS-3, RS-1, PR-3, PR-2, PR-1, SP-6, SP-4, SP-3, RJ-1, ES, MG-2, BA-2 MG-3 RS-5, RS-4, RS-3, RS-2, SC-1, PR-3, PR-2, SP-4, RJ-2, MG-3, MG-2, MG-1, BA-2 PR-1, SP-6, SP-4 RJ-1 RJ-1 PR-1, SP-6, SP-5, RJ-1 RJ-1, ES, MG-2 BA-1 PR-1, SP-6, SP-5, SP-4, SP-3, RJ-1, MG-2 SP-3, BA-2, ES BA-4, BA-3, BA-2 SP-6, SP-5, SP-3, BA-4, BA-3, BA-2, BA-1 SP-6, SP-3 T. sprengeliana Klotzsch ex Mez T. stricta Sol. ex Sims T. tenuifolia L. T. tricholepis Baker T. usneoides (L.) L. Vriesea altodaserrae L.B. Sm. V. altomacaensis A. Costa V. arachnoidea A. Costa V. atra Mez V. bituminosa Wawra V. blackburniana Leme V. carinata Wawra V. drepanocarpa (Baker) Mez V. duvaliana E. Morren V. ensiformis (Vell.) Beer V. erythrodactylon (E. Morren) E. Morren ex Mez V. flammea L.B. Sm. V. friburgensis Mez V. gamba F. Mull. V. gigantea Mart. ex Schult. f. V. gracilior (L.B. Sm.) Leme V. guttata Linden & André V. heterostachys (Baker) L.B. Sm. V. hieroglyphica (Carrière) E. Morren V. hydrophora Ule V. incurvata Gaudich. V. inflata (Wawra) Wawra V. longicaulis (Baker) Mez V. longiscapa Ule V. minuta Leme V. modesta Mez V. paludosa L.B. Sm. V. paratiensis E. Pereira V. pauperrima E. Pereira V. penduliflora L.B. Sm. V. philippocoburgii Wawra V. platynema Gaudich. V. platzmannii E. Morren V. procera (Mart. ex Schult. f.) Wittm. V. psittacina (Hook.) Lindl. V. regina (Vell.) Beer V. reitzii Leme & A. Costa V. rhodostachys L.B. Sm. V. rodigasiana E. Morren V. ruschii L.B. Sm. V. scalaris E. Morren V. simplex (Vell.) Beer V. sparsiflora L.B. Sm. V. sucrei L.B. Sm. & Read V. tijucana E. Pereira V. triligulata Mez V. tucumanensis Mez V. vagans (L.B. Sm.) L.B. Sm. Witrrockia gigantea (Baker) Leme SC-1, PR-1, SP-6, SP-5, SP-4, RJ-1, BA-4, BA-3, BA-2 RS-4, RS-3, RS-1, PR-3, PR-1, SP-4, SP-3, MG-2 BA-3 RS-4, RS-3, PR-1, SP-6, SP-5, SP-4, MG-1 ES, MG-3, MG-1, BA-2 MG-2 SP-6, RJ-1, MG-2 SP-4, RJ-1 SP-4, RJ-1 SC-1, SP-6, SP-5, SP-4, SP-3 SP-6, SP-3, RJ-1 RJ-1, ES, MG-2, BA-3 SP-3, RJ-1, BA-3 BA-4, BA-2, BA-1 MG-3 SP-6 RJ-1, MG-3, BA-4, BA-3 MG-1 MG-2 RS-4, SC-1, PR-1, SP-5, SP-4, SP-3, RJ-1 PR-3, SP-4, BA-2 RS-4, PR-1, SP-6 RS-3, PR-1, RJ-2, BA-2 RS-3, ES, BA-2, BA-1 BA-3 PR-2 BA-1 PR-1, SP-6, SP-3, BA-3 BA-3, BA-4 SP-3, MG-1 ES, BA-4, BA-2, BA-1 RJ-1 RJ-2 BA-2 RJ-1 RS-5 RS-4, SC-1, PR-1, SP-6, SP-5, SP-4, SP-3, RJ-1, MG-3 MG-2 671 References Alves, R.J.V., Kolbek, J., Becker, J., 2008. Vascular epiphyte vegetation in rocky savannas of southeastern Brazil. Nord. J. Bot. 26, 101–117. Amorim, A.M., Fiaschi, P., Jardim, J.G., Thomas, W.W., Clifton, B.C., Carvalho, A.M.V., 2005. The vascular plants of a forest fragment in southeastern Bahia, Brazil. Sida 21, 1727–1752. Amorim, A.M.A., Thomas, W.W., Carvalho, A.M.V., Jardim, J.G., 2008. Floristics of the Una Biological Reserve, Bahia, Brasil. Mem. New York Bot. Gard. 100, 67–146. Amorim, A.M., Jardim, J.G., Lopes, M.M.M., Fiaschi, P., Borges, R.A., Perdiz, R.O., Thomas, W.W., 2009. Angiospermas em remanescentes de floresta montana no sul da Bahia, Brasil. Biota Neotrop. 9, 313–348. Barfuss, M.H.J., Samuel, R., Till, W., Stuessy, T.F., 2005. Phylogenetic relationships in subfamily Tillandsioideae (Bromeliaceae) based on DNA sequence data from seven plastid regions. Am. J. Bot. 92, 337–351. Benzing, D.H., 1990. Vascular Epiphytes. General Biology and Related Biota. Cambridge University Press, Cambridge. Benzing, D.H., 2000. Bromeliaceae. Profile of an adaptative radiation. Cambridge University Press, Cambridge. Bonnett, A., Queiroz, M.H., 2006. Estratificação vertical de bromélias epifíticas em diferentes estádios sucessionais da Floresta Ombrófila Densa, Ilha de Santa Catarina, Santa Catarina, Brasil. Rev. Bras. Bot. 29, 217–228. Breier, T.B., 2005. O epifitismo vascular em florestas do sudeste do Brasil. PhD Thesis. Universidade Estadual de Campinas, Campinas. Buzatto, C.R., Severo, B.M.A., Waechter, J.L., 2008. Composição florística e distribuição ecológica de epífitos vasculares na Floresta Nacional de Passo Fundo, Rio Grande do Sul. Iheringia Ser. Bot. 63, 231–239. Cody, M.L., 1993. Bird diversity components within and between habitats in Australia. In: Ricklefs, R.E., Schluter, D. (Eds.), Species Diversity in Ecological Communities. Historical and Geographical Perspectives. Chicago Press, Chicago, pp. 147–158. Costa, A.F., Wendt, T., 2007. Bromeliaceae na região de Macaé de Cima, Nova Friburgo, Rio de Janeiro, Brasil. Rodriguésia 58, 905–939. Fiaschi, P., Pirani, J.R., 2009. Review of plant biogeographic studies in Brazil. J. Syst. Evol. 47, 477–496. Fischer, A.E., Araújo, A.C., 1995. Spatial organization of a bromeliad community in the Atlantic rainforest, south-eastern Brazil. J. Trop. Ecol. 11, 559–567. Fontoura, T., Reinert, F., 2009. Habitat utilization and CAM occurrence among epiphytic bromeliads in a dry forest from Southeastern Brazil. Rev. Bras. Bot. 32, 519–528. Fontoura, T., Santos, F.A.M., 2010. Geographic distribution of epiphytic bromeliads of the Una region, Northeastern Brazil. Biota Neotrop. 10, 127–132. Fontoura, T., Costa, A., Wendt, T., 1991. Preliminary checklist of the Bromeliaceae of Rio de Janeiro State, Brazil. Selbyana 12, 5–45. Fontoura, T., Sylvestre, L.S., Vaz, A.M.F., Vieira, C.M., 1997. Epífitas vasculares, hemiepífitas e hemiparasitas da Reserva Ecológica de Macaé de Cima. In: Lima, H.C., Guedes-Bruni, R.R. (Eds.), Serra de Macaé de Cima: diversidade florística e conservação em Mata Atlântica. Jardim Botânico do Rio de Janeiro, Rio de Janeiro, pp. 89–102. Gentry, A., Dodson, C.H., 1987. Diversity and biogeography of Neotropical vascular epiphytes. Ann. Missouri Bot. Gard. 74, 205–233. Giongo, C., Waechter, J.L., 2004. Composição florística e estrutura comunitária de epífitos vasculares em uma floresta de galeria na Depressão Central do Rio Grande do Sul. Rev. Bras. Bot. 27, 563–572. Givnish, T.J., Millam, K.C., Berry, P.E., Sytsma, K.J., 2007. Phylogeny, adaptive radiation, and historical biogeography of Bromeliaceae inferred from ndhF sequence data. In: Columbus, J.T., Friar, E.A., Porter, J.M., Prince, L.M., Simpson, M.G. (Eds.), Monocots: Comparative Biology and Evolution – Poales. Rancho Santa Ana Botanic Garden, Claremont, pp. 3–26. Givnish, T.J., et al., 2011. Phylogeny, adaptive radiation, and historical biogeography in Bromeliaceae: insights from an eight-locus plastid phylogeny. Am. J. Bot. 98, 872–895. Gonçalves, C.N., Waechter, J.L., 2002. Epífitos vasculares sobre espécimes de Ficus organensis isolados no norte da planície costeira do Rio Grande do Sul: padrões de abundância e distribuição. Acta Bot. Bras. 16, 429–441. Gonçalves, C.N., Waechter, J.L., 2003. Aspectos florísticos e ecológicos de epífitos vasculares sobre figueiras isoladas no norte da planície costeira do Rio Grande do Sul. Acta Bot. Bras. 17, 89–100. Haffer, J., 1987. Quaternary history of Tropical America. In: Whitmore, T.C., Prance, G.T. (Eds.), Biogeography and Quaternary History in Tropical America. Clarendon Press, Oxford, pp. 1–16. Hertel, R.J.G., 1949. Contribuição à ecologia da Serra do Mar (vertente oeste) do Paraná. Tese de Livre Docência. Universidade do Paraná, Curitiba. Ibisch, P.L., Nowicki, C., Vasquez, R., 2001. Towards an understanding of diversity patterns and conservation requirements of the Bolivian Bromeliaceae. J. Bromeliad Soc. 51, 99–112. Jones, H.G., 1992. Plants and Microclimate, 2nd ed. Cambridge University Press, Cambridge. Kersten, R.A., 2010. Epífitas vasculares – Histórico, participação taxonômica e aspectos relevantes, com ênfase na Mata Atlântica. Hoehnea 37, 9–38. Kersten, R.A., Silva, M.S., 2001. Composição florística e estrutura do componente epifítico vascular em floresta de planície litorânea na Ilha do Mel, Paraná, Brasil. Rev. Bras. Bot. 24, 213–226. 672 T. Fontoura et al. / Flora 207 (2012) 662–672 Kersten, R.A., Silva, S.M., 2002. Florística e estrutura do componente epifítico vascular em Floresta Ombrófila Mista Aluvial do rio Barigüi, Paraná, Brasil. Rev. Bras. Bot. 25, 259–267. Kersten, R.A., Kuniyoshi, Y.S., Roderjan, C.V., 2009. Epífitas vasculares em duas formações ribeirinhas adjacentes na bacia do rio Iguaçu – Terceiro Planalto Paranaense. Iheringia Ser. Bot. 64, 33–43. Kessler, M., 2000. Elevational gradients in species richness and endemism of selected plant groups in the central Bolivian Andes. Plant Ecol. 149, 181–193. Kessler, M., 2001. Patterns of diversity and range size of selected plant groups along an elevational transect in the Bolivian Andes. Biodivers. Conserv. 10, 1897–1920. Kessler, M., 2002a. Environmental patterns and ecological correlates of rangesize among bromeliad communities of Andean forests in Bolívia. Bot. Rev. 68, 100–127. Kessler, M., 2002b. Species richness and ecophysiological type among Bolivian bromeliad communities. Biodivers. Conserv. 11, 987–1010. Küpper, W., Kreft, H., Nieder, J., Köster, N., Barthlott, W., 2004. Large-scale diversity patterns of vascular epiphytes in Neotropical montane rain forests. J. Biogeogr. 31, 1477–1487. Krömer, T., Kessler, M., 2005. Diversity patterns of vascular epiphytes along an elevational gradient in the Andes. J. Biogeogr. 32, 1799–1809. Laube, S., Zotz, G., 2006. Neither host-specific nor random: vascular epiphytes on three tree species in a Panamanian Lowland Forest. Ann. Bot. 97, 1103–1114. Leme, E.M.C., Siqueira-Filho, J.A., 2006a. Taxonomia das bromélias dos fragmentos de Mata Atlântica de Pernambuco e Alagoas. In: Siqueira-Filho, J.A., Leme, E.M.C. (Eds.), Fragmentos de Mata Atlântica do nordeste. Biodiversidade, conservação e suas bromélias. Andrea Jakobsson Estúdio, Rio de Janeiro, pp. 191–382. Leme, E.M.C., Siqueira-Filho, J.A., 2006b. Considerações taxonômicas correlacionadas às bromélias de Pernambuco e Alagoas. In: Siqueira-Filho, J.A., Leme, E.M.C. (Eds.), Fragmentos de Mata Atlântica do nordeste. Biodiversidade, conservação e suas bromélias. Andrea Jakobsson Estúdio, Rio de Janeiro, pp. 383–407. Marques, M.C.M., Swaine, M.D., Liebsch, D., 2010. Diversity, distribution, and floristic differentiation of the coastal lowland vegetation: implications for the conservation of the Brazilian Atlantic Forest. Biodivers. Conserv. 20, 153–168. Martinelli, G., Vieira, C.M., Gonzalez, M., Leitman, P., Piratininga, A., Costa, A.F., Forzza, R.C., 2008. Bromeliaceae da Mata Atlântica brasileira: lista de espécies, distribuição e conservação. Rodriguésia 59, 209–258. McCune, B., Mefford, M.J., 1997. Multivariate Analysis of Ecological Data, ver. 3.11. MjM Software, Oregon. Menini-Neto, L., Forzza, R.C., Zappi, D., 2009a. Angiosperm epiphytes as conservation indicators in forest fragments: a case study from southeastern Minas Gerais, Brazil. Biodivers. Conserv. 18, 3785–3807. Menini-Neto, L., et al., 2009b. Flora vascular não-arbórea de uma floresta de grota na Serra da Mantiqueira, Zona da Mata de Minas Gerais, Brasil. Biota Neotrop. 9, 149–161. Oliveira, M.L.A.A., Balbueno, R.A., Senna, R.M., 2005. Levantamento florístico de fragmentos florestais na bacia hidrográfica do rio Gravataí, Rio Grande do Sul, Brasil. Iheringia Ser. Bot. 60, 269–284. Oliveira-Filho, A.T., Fontes, M.A., 2000. Patterns of floristic differentiation among Atlantic Forests in Southeastern Brazil and the influence of climate. Biotropica 32, 793–810. Oliveira-Filho, A.T., et al., 2005. Análise florística do compartimento arbóreo de áreas de Floresta Atlântica sensu lato na região das bacias do leste (Bahia, Minas Gerais, Espírito Santo e Rio de Janeiro). Rodriguésia 56, 185–235. Pontes, R.A.S., Agra, M.F., 2006. Flora da Paraíba, Brasil: Tillandsia L. (Bromeliaceae). Rodriguésia 57, 47–61. Reis, J.R.M., Fontoura, T., 2009. Diversidade de bromélias epífitas na Reserva Particular do Patrimônio Natural Serra do Teimoso – Jussari, BA. Biota Neotrop. 9, 73–79. Rizzini, C.T., 1992. Tratado de fitogeografia do Brasil, 2a . ed. Âmbito Cultura Edições Ltda, Rio de Janeiro. Rogalski, J.M., Zanin, E.M., 2003. Composição florística de epífitos vasculares no estreito de Augusto César, Floresta Estacional Decidual do Rio Uruguai, RS, Brasil. Rev. Bras. Bot. 26, 551–556. Safford, H.D., 2007. Brazilian Páramos IV. Phytogeography of the campos de altitude. J. Biogeogr. 34, 1701–1722. Santos, C.G., 2000. Distribuição espacial, fenologia e polinização de Bromeliaceae na Mata Atlântica do alto da Serra de Paranapiacaba. PhD Thesis. Universidade Estadual de Campinas, Campinas. Sass, C., Specht, C.D., 2010. Phylogenetic estimation of the core Bromelioids with an emphasis on the genus Aechmea (Bromeliaceae). Mol. Phylogenet. Evol. 55, 559–571. Schulte, K., Horres, R., Zizka, G., 2005. Molecular phylogeny of Bromeliaceae and its implications on biogeography and evolution of CAM in the family (Poales, Bromeliaceae). Senckenbergiana Biol. 85, 113–125. Schulte, K., Barfuss, M.H.J., Zizka, G., 2009. Phylogeny of Bromelioideae (Bromeliaceae) inferred from nuclear and plastid DNA loci reveals the evolution of the tank habit within the subfamily. Mol. Phylogenet. Evol. 51, 327–339. Scudeller, V.V., Martins, F.R., Shepherd, J.G., 2001. Distribution and abundance of arboreal species in the Atlantic Ombrophilous Dense Forest in Southeastern Brazil. Plant Ecol. 152, 185–199. Shepherd, J., 1995. Fitopac. Manual do Usuário. Universidade Estadual de Campinas, Campinas. Smith, L.B., Downs, R.J., 1974. Pitcairnioideae (Bromeliaceae). Fl. Neotrop. Monogr. 14, 1–658. Smith, L.B., Downs, R.J., 1977. Bromeliaceae (Tillandsioideae). Fl. Neotrop. Monogr. 14, 663–1492. Sneath, P.H.A., Sokal, R.R., 1973. Numerical Taxonomy. The Principles and Practice of Numerical Classification. W.H. Freeman and Company, San Francisco. Sousa, G.M., Wanderley, M.G.L., 2000. Aechmea Ruiz & Pav. (Bromeliaceae) do estado de Pernambuco, Brasil. Acta Bot. Bras. 14, 77–97. Stehmann, J.R., Forzza, R.C., Sobral, M., Kamino, L.H.Y., 2009. Gimnospermas e Angiospermas. In: Stehmann, J.R., Forzza, R.C., Salino, A., Sobral, M., Costa, D.P., Kamino, L.H.Y. (Eds.), Plantas da Floresta Atlântica. Jardim Botânico do Rio de Janeiro, Rio de Janeiro, pp. 27–40. Ter Braak, C.F.J., 1997. Ordination. In: Jongman, R.H.G., Ter Braak, C.J.F., van Tongeren, O.F.R. (Eds.), Data Analysis in Community and Landscape Ecology. Cambridge University Press, New York, pp. 91–173. Varassin, I.G., 2002. Estrutura especial e temporal de uma comunidade de Bromeliaceae e seus polinizadores em Floresta Atlântica no sudeste do Brasil. PhD Thesis. Universidade Estadual de Campinas, Campinas. Versieux, L.M., Wendt, T., 2006. Checklist of Bromeliaceae of Minas Gerais, Brazil, with notes on taxonomy and endemism. Selbyana 27, 107–146. Versieux, L.M., Wendt, T., 2007. Bromeliaceae diversity and conservation in Minas Gerais, Brazil. Biodivers. Conserv. 16, 2989–3009. Waechter, J.L., 1998. Epifitismo vascular em uma floresta de restinga do Brasil subtropical. Ciênc. Nat. 20, 43–66. Wanderley, M.G.L., Martins, S.E., 2007. Bromeliaceae: coordenação, descrição da família e chave de gêneros. In: Wanderley, M.G.L., Shepherd, G.J., Melhem, T.S., Giulietti, A.M. (Eds.), Flora Fanerogâmica do Estado de São Paulo, vol. 5. FAPESP, Instituto de Botânica, São Paulo, pp. 39–161. Whitmore, T.C., Prance, G.T., 1987. Biogeography and Quaternary History of Tropical America. Clarendon Press, Oxford. Wolf, J.H.D., Flamenco-S, A., 2003. Patterns in species richness and distribution of vascular epiphytes in Chiapas, México. J. Biogeogr. 30, 1689–1707.