Abiotic and vertebrate seed dispersal in the Brazilian Atlantic forest: implications for forest regeneration

Abiotic and vertebrate seed dispersal in the Brazilian Atlantic forest: implications for forest regeneration

Biological Conservation 106 (2002) 165–176 www.elsevier.com/locate/biocon Abiotic and vertebrate seed dispersal in the Brazilian Atlantic forest: imp...

193KB Sizes 1 Downloads 106 Views

Biological Conservation 106 (2002) 165–176 www.elsevier.com/locate/biocon

Abiotic and vertebrate seed dispersal in the Brazilian Atlantic forest: implications for forest regeneration Marcelo Tabarellia,*, Carlos A. Peresb a

Departamento de Botaˆnica—CCB, Universidade Federal de Pernambuco, Recife, PE 50670-901, Brazil b School of Environmental Sciences, University of East Anglia, Norwich NR4 7TJ, UK Received 7 January 2001; received in revised form 30 July 2001; accepted 19 October 2001

Abstract This study documents the proportion of woody plant species having different modes of seed dispersal, and classes of seed and fruit size at 16 Atlantic montane forest plots of southeastern Brazil (23–24 S). These plots represent six chronosequences (from 5year-old to old-growth forest) of forest regeneration following small-scale shifting agriculture. Our results indicate that there is a gradual but predictable increase in the number of woody plant species relying on vertebrate-mediated seed dispersal in increasingly older plots. Moreover, the percentage of plant species bearing small seeds and fruits ( < 0.6 cm in length) was more than halved from earlier to older forest plots, while the percentage of plant species producing middle-sized seeds and fruits at least doubled. Plant species belonging to the Melastomataceae, Myrsinaceae, Rubiaceae and Flacourtiaceae comprised a large proportion of the small-seeded and fruit species (33–53%), and largely occurred in early successional forest plots (5–30 years old). In contrast, species belonging to the Myrtaceae and Lauraceae represented most of the species (51–72%) bearing medium-sized seeds and fruits (0.6–1.5 cm in length), and occurred primarily in the old-growth forests. Shifts in the relative importance of dispersal strategies during the regeneration process of Atlantic montane forest, as well as in diaspore size appear to be related to the balance between early successional and shade-tolerant species associated with particular life forms and plant families. Finally, we discuss the relationships between the species richness of Myrtaceae and Lauraceae trees, seed dispersal by medium to large vertebrates, and possible regeneration scenarios for the Atlantic forest. # 2002 Elsevier Science Ltd. All rights reserved. Keywords: Brazilian Atlantic forest; Forest regeneration; Frugivorous vertebrates; Seed dispersal; Species richness

1. Introduction Animal-mediated seed dispersal is usually the most important mechanism of propagule dissemination in tropical forests where a vast proportion of the woody plant species are dispersed by vertebrates rather than wind, water, or other abiotic processes (e.g. van Roosmalen, 1985; Charles-Dominique, 1993). Although seeds of most tropical forest plants are dispersed by vertebrates, the relative importance of different dispersal mechanisms is considerably different between neotropical and paleotropical forests (Howe and Smallwood, 1982; Richards, 1996), evergreen and deciduous forests (Gentry, 1995), and early successional and mature forests (Brown and Lugo, 1990). In general, the proportion of vertebrate-dispersed species is greater in

* Corresponding author.

wet, late-successional or mature forests (Howe and Smallwood, 1982). This is particularly the case of species-rich neotropical forests, where as many as 87–90% of the woody plant species in a local flora produces diaspores that are gut-dispersed by vertebrates (Peres and van Roosmalen, 2002). Previous studies in Amazonian forests (Uhl, 1987; Nepstad et al., 1996; Vieira et al., 1996) and the Brazilian Atlantic forest (Tabarelli and Mantovani, 1999a) suggest that tree and shrubs species dispersed by abiotic means are gradually replaced by those dispersed by vertebrates throughout the forest regeneration process. Among the vertebratedispersed species, those in early secondary forests producing small fruits and seeds appear to be gradually replaced by those producing medium to large fruits and seeds in increasingly older forests (Howe and Smallwood, 1982; Richards, 1996; Vieira et al., 1996). This gradual increase in the species richness and abundance of vertebrate-dispersed woody plant species

0006-3207/02/$ - see front matter # 2002 Elsevier Science Ltd. All rights reserved. PII: S0006-3207(01)00243-9

166

M. Tabarelli, C.A. Peres / Biological Conservation 106 (2002) 165–176

from early to late-successional forests reinforces the pantropical notion that forest regeneration processes that are able to preserve most, if not all, plant species in the source flora rely heavily on effective seed dispersal by frugivorous vertebrates (cf. Nepstad et al., 1996; Vieira et al., 1996; Corlett, 1998; Guaricata and Pinard, 1998; Duncan and Chapman, 1999). For example, natural succession in small, degraded forest fragments in Singapore is unlikely to restore the original plant species composition (Corlett and Turner, 1997). Large-seeded forest trees are simply unable to colonise new patches of secondary forests because their large-bodied vertebrate dispersers are now locally extinct. Birds in Indomalayan fragmented forest landscapes typically have gape widths of 10–15 mm, and are thus too small to ingest whole seeds of most forest fruits. Bird species with larger gape widths and large mammals capable of dispersing largeseeded fruits are highly vulnerable to selective hunting, fragmentation and habitat loss (Corlett, 1998). Likewise, in northern Thailand, seedling recruitment rates of two vertebrate-dispersed plant species in abandoned agricultural clearings are severely limited by low rates of seed dispersal (Hardwick et al., 1997). In the island of La Re´union, large-seeded (> 20 mm) fleshy-fruit species show very low colonisation rates and are missing from young lava flows (The´baud and Strasberg, 1997), perhaps because of the extremely impoverished frugivore fauna following an alarming number of large-vertebrate extinctions in the last three centuries (Cox et al., 1991). It remains to be seen, however, whether or not largeseeded plants can adjust to the loss of key seed-dispersal services. The Brazilian Atlantic forest is one of the highest global priorities for conservation (Myers et al., 2000), with only 7% of the original pre-Columbian forest cover of over 1.2106 km2 still standing (Brown and Brown, 1992; Myers et al., 2000). Most of the remaining forest was retained within private properties (Viana et al., 1997), and continues to be disturbed by a combination of shifting agriculture, logging and wildfires (Oliveira et al., 1994) as well as degradation of small forest fragments (Gascon et al., 2000). Human disturbance typically creates a vegetation mosaic where early and late-successional forests are converted to subsistence and cash crops and later abandoned to natural forest succession (Tabarelli and Mantovani, 1999a). Over large tracts of the remaining Atlantic forest, large-bodied frugivorous vertebrates, such as primates, toucans, cracids, and large cotingas, have been extirpated or significantly reduced in numbers (Almeida et al., 1995; Galetti et al., 1997; Silva and Tabarelli, 2000). It is therefore possible that background rates of vertebrate-mediated seed delivery have been significantly reduced thus delaying, curbing, or precluding forest regeneration so that late-successional forests will never be able to attain the full complement of plant species typical of original old growth forests.

Here we examine the proportion of woody plant species occurring within different modes of seed dispersal, and classes of fruit and seed size within 16 Atlantic montane forest plots (comprising six sites) of southeastern Brazil at different stages of regeneration. These 16 plots represent six chronosequences (ranging from 5year-old to old-growth forest) of forest regeneration following clear-cuts in shifting agriculture plots. We examine three questions on the relationship between seed dispersal and forest regeneration: (1) are wind and water-dispersed species replaced by those dispersed by vertebrates in the course of the forest regeneration process? (2) are small-seeded species gradually replaced by large-seeded counterparts? and (3) are small-fruit species gradually replaced by those producing large fruits? The relationships between seed dispersal mode, regeneration strategy, plant life form, and taxonomic groupings are also discussed. Finally, we discuss the relationships between the species richness of Myrtaceae and Lauraceae trees, seed dispersal by medium to large vertebrates, and possible regeneration scenarios for old secondary stands of the Brazilian Atlantic forest regarding woody plant species richness and composition.

2. Methods 2.1. Forest chronosequences We selected a set of published plant inventories conducted in 16 Atlantic forest plots of varying regeneration stages or ages, located at six different sites (Fig. 1). All sites are located between 1 and 100 km inland from the coastline and along the slopes of the Serra do Mar kniferidge within three neighbouring states of southeastern Brazil: Rio de Janeiro, Sa˜o Paulo, and Parana´ (23–24 S; Table 1). According to the Koeppen’s classification system, the regional climate is humid-temperate, lacking a demarcated dry season. Mean rainfall at all study sites ranged from 1400 to 2767 mm year1, with the wettest and the driest periods of the year in December–March, and July–August, respectively. The predominant soil types at all sites are lithosols and podzols (Radambrasil, 1983). The vegetation is best described as tropical montane rain forest at elevations of 100–1100 m.a.s.l., which comprise one of the major Atlantic forest types (submontane and montane categories of Veloso et al., 1991). Detailed descriptions of the plant species composition in these forests can be found elsewhere (Melo and Mantovani, 1994; GuedesBruni et al., 1997). The 16 forest plots considered here were located either within (13 plots in five sites: Cubata˜o, Intervales, Iporanga, Macae´ de Cima, Santa Virgı´nia ) or along the boundaries of protected areas (three plots in one site: Morretes; Table 1), which still harbour large patches of

167

M. Tabarelli, C.A. Peres / Biological Conservation 106 (2002) 165–176

Fig. 1. Location of the 16 forest plots (within six sites) sampled in the montane Atlantic forest of southeastern Brazil.

Table 1 Main characteristics of the six Atlantic forest sites surveyed along the Serra do Mar, States of Rio de Janeiro, Sa˜o Paulo and Parana´, southeastern Brazil Site

Location

Cubata˜o

23 220 S, 45 260 W

100

2767

Intervales

24 120 S, 48 030 W

900

1720

Iporanga

24 300 S, 48 300 W

200–250

1400–1700

Macae´ de Cima Morretes

22 210 S, 42 270 W

1000–1100

2128

25 300 S, 49 000 W

50–400

2290

23 170 S, 45 300 W

900

2140

Santa Virgı´nia a

Elevation (m)

Rainfall (mm/year)

Landscape contexta

Data sources

Protected forests of Serra do Mar State Park (360,000 ha) Protected forests of Intervales State Park (36,000 ha) Protected forests of Petar State Park (35,000 ha) Protected forests of Macae´ de Cima Biological Reserve (7200 ha) Protected forests of Serra do Mar (386,000 ha) Protected forests of Santa Virgı´nia Reserve (5000 ha)

Leita˜o Filho (1993) Nascimento (1994) Torezan (1995) Pessoa et al. (1997) Guedes-Bruni et al. (1997) Guapyassu (1994) Tabarelli (1997) Tabarelli and Mantovani (1999a)

Refers to the predominant surrounding landscape matrix in which the forest sites are located (Lima and Capobianco, 1997).

continuous forest with their typical plant and vertebrate species composition. The smallest of these protected areas—Macae´ de Cima Biological Reserve—encompasses 7200 ha of Atlantic montane forest surrounded by 35,000 ha of continuos forest (Lima and GuedesBruni, 1997). The largest one—Serra do Mar State Park—protects 360,000 ha of continuous forest, including the Santa Virgı´nia Reserve (Lima and Capobianco, 1997). These 16 plots represent six chronosequences of forest plots left to regenerate (ranging from 5-yearold to old-growth forest) following slash-and-burn

agriculture. These were selected for this analysis based on the best available floristic data and information on past history of land use (Table 2). Prior to floristic surveys, the primary forest at all regenerating forest plots had been clear-felled, burned and used for subsistence crops for 2–10 years, and subsequently abandoned to revert to secondary forest. The plots ranged from 1 to 5 ha in size, according to detailed site descriptions provided by the authors of the different studies which were conducted on the basis of different methods. However, plant sampling criteria used in these studies [from 6 cm

168

M. Tabarelli, C.A. Peres / Biological Conservation 106 (2002) 165–176

Table 2 Sampling effort and criteria used during the floristic inventories conducted at the six Atlantic forest sites considered in this study Site

No. of forest plots

Age of forest plots (year)

Sample sizea

Sampling criteriab

No. of woody plant speciesc

Cubata˜o Intervales Iporanga Macae´ de Cima Morretes Santa Virgı´nia

2 3 3 2 3 3

15, old growth 5, 13, 18 5, 15, 50 30, old growth 13, 25,old growth 18, 40,old growth

370 792 – 4306 1126 1672

56.3 cm in DBH 0.95 cm diameter at soil level >10 cm in height 55 cm in DBH 56.3 cm in DBH 53.6 cm in DBH

61, 75 17, 29, 59 26, 41, 93 157, 185 23, 41, 77 56, 90, 84

a b c

Refers to total number of individuals sampled at each site. Refers to the minimum size (diameter at breast height, DBH, or height) of plants sampled. Number of woody plant species sampled (ordered from the youngest to the oldest forest plot).

diameter at breast height (DBH) to all woody plants taller than 1 m] were considered to be appropriate to record a representative set of tree and shrub species occurring at each plot.

inspection of herbarium specimens at the herbaria of Instituto de Botaˆnica de Sa˜o Paulo, Brazil.

2.2. Seed dispersal mode

In order to examine the relationships between seed dispersal mode (vertebrate vs. abiotic dispersal), seed and fruit size, plant species composition and stage of forest succession, we quantified the number (and proportion) of species within the Melastomataceae, Rubiaceae, Flacourtiaceae, Myrsinaceae, Lauraceae and Myrtaceae in early secondary (5–30 years old) and old-growth forests. These six families were selected because (1) they are widely distributed in the Atlantic montane forest, (2) they often account for over half of the woody plant species richness in plots of this forest, and (3) there are vast records on literature about the natural history of their species (Klein, 1980; Silva and Leita˜o Filho, 1982; Mori et al., 1983; Lima and Guedes-Bruni, 1997; Tabarelli, 1997; Tabarelli and Mantovani, 1999b).

All woody plant species sampled within each plot were classified according to two mutually exclusive modes: (1) vertebrate-dispersed species, or those producing diaspores adhered to a fleshy pulp, aril, or other features typically associated with vertebrate dispersal agents; and (2) species dispersed by abiotic means, or those presenting winged seeds, plumes, or other winddispersal devices that slow the rate of seed fall, or those dispersed entirely by free fall or propelled explosively by a fruit that bursts suddenly open (van der Pijl, 1982). Vertebrate-dispersed species were further classified in relation to seed size according to the following sizeclasses: (1) seeds smaller than 0.6 cm in length; (2) seeds sizing 0.6–1.5 cm; (3) seeds sizing 1.6–3.0 cm; and (4) seeds larger than 3.0 cm in length. Vertebrate-dispersed species were also classified in relation to fruit size according to the same categories described above. ‘‘We use these categories because there is a great deal of intraspecific variation in fruit and seed size within these size classes for a large proportion of the woody plant species in the Atlantic forest (Reitz, 1965; Barroso et al., 1999; Silva and Tabarelli, 2000; M. Tabarelli, unpublished data). Given the nature of the flora in the Atlantic forest, these size-classes are assumed to represent natural categories, although this has never been adequately tested. For convenience, we simply refer to seeds and fruits within these increasingly larger size-classes as small, medium-sized, large and very large. Classification of plant species according to seeddispersal mode and seed and fruit size classes was based on (1) our own knowledge of the flora sampled in relation to fruit and seed morphology; (2) detailed accounts of species life history traits available in the literature (e.g. Flora Neotropica, Reitz, 1965; van Roosmalen, 1985; Lorenzi, 1998; Barroso et al., 1999); and (3)

2.3. Plant species composition

2.4. Data analysis We examined the relationship between age of forest plots and the proportion of species with different dispersal modes using Spearman rank correlations (Siegel, 1979). For each site, between-plot differences in the proportion of species in different dispersal modes were examined using log-likelihood ratio (G) tests (Sokal and Rohlf, 1995). We used the same procedures to examine the relationship between age of forest plots and the relative species richness of different size classes of fruits and seeds. In the absence of precise historical records on the oldgrowth forest plots sampled, these were assumed to be approximately 120 yrs old, given the history of deforestation and forest fragmentation in the region and the minimum amount of time required for complete regeneration of Atlantic montane forests following shifting agriculture (Klein, 1980). G-tests were also used to examine differences in the frequency of species occurrence of Melastomataceae, Rubiaceae, Flacourtiaceae,

M. Tabarelli, C.A. Peres / Biological Conservation 106 (2002) 165–176

Myrsinaceae, Lauraceae and Myrtaceae between the early secondary (5–30 years old) and old-growth forests. All analyses were performed using SYSTAT 6.0 (Wilkinson, 1996). We also applied the percentage of species within these six families in the 16 plots to a detrended correspondence analysis (DCA; sensu Hill and Gauch, 1980), using CANOCO version 3.11.5 (ter Braak, 1987–1992). If the percentage of plant species within these families varied linearly from younger to older forests, plots would be expected to cluster in the DCA primarily according to their age, rather than geographic region, physical proximity, or elevation. In most of the analyses we compare plots both within and among different sites. At the first spatial scale, between-plot differences are expected to be correlated only with the age of forest plots since sampling method, sampling effort and overall plot size were identical. At the second level of comparisons, we were unable to control for some differences in the surrounding vegetation matrix, plot size, sample size and sampling criteria, which could introduce some bias in our results. However, such differences are expected to explain any trend detected only if differences in the surrounding matrix, plot size, sample size and sampling criteria varied linearly across the age of forest plots which was decidedly not the case. We therefore assume that between-plot differences regarding dispersal mode and percentage of species within categories of seed and fruit size are primarily associated with forest age rather than particular differences between sites, plot size or methods. If this is

169

true than the trends uncovered should share similar causal relationships at both spatial scales.

3. Results 3.1. Abiotic vs. vertebrate seed dispersal Vertebrate seed dispersal mode prevailed in all 16 forest plots. However, this was highly variable across sites ranging from 52.9% of the species in a 5-year-old plot at Intervales to 98.7% in an old growth plot at Santa Virgı´nia. There was a strong positive correlation between the age of forest plots and the percentage of vertebrate-dispersed plant species within them (rs=0.84, n=16, P < 0.001; Fig. 2). This correlation remained significant after the four old-growth plots were discarded from the analysis (rs=0.78, n=12, P < 0.001), suggesting that there was a trend throughout different stages of forest succession. At three different sites (Iporanga, Macae´ and Santa Virgı´nia) the proportion of vertebrate-dispersed species significantly increased from the youngest to the oldest forest plots (G tests; P < 0.05 in all possible pairwise comparisons). 3.2. Small vs. medium-sized seeds and fruits Small (< 0.6 cm) and medium-seeded (0.6–1.5 cm) plant species prevailed in all forest plots, ranging from 24.6% to 72.2% (small-seeded) and from 14.3 to 59.2% (medium-seeded) of the species sampled (Fig. 3). Large

Fig. 2. Relationship between the successional stage of 16 montane Atlantic forest plots in terms of recovery (in years) since the last clear-cut, and the percentage of plant species in each woody flora known to be dispersed by frugivorous vertebrates of any size.

170

M. Tabarelli, C.A. Peres / Biological Conservation 106 (2002) 165–176

Fig. 3. Relationship between the age of 16 montane Atlantic forest plots and the percentage of woody plant species in different classes of seed size.

and very large-seeded species thus represented a relatively small percentage of all the species sampled in the floras of the 16 plots (< 25%). We also found a strong negative correlation between the age of forest plots and the percentage of small-seeded plants as it was reduced by three thirds throughout the successional process (rs=0.83, n=16, P=0.0001). In contrast, the percentage of medium-seeded plant species was positively correlated with the age of forest plots (rs=0.85, n=16, P=0.0001). Mirroring the pattern shown for seed dispersal mode, these correlations remained highly significant even after old-growth plots were removed from the analyses (rs=0.73, n=12, P=0.006; rs=0.78, n=12, P=0.002). A similar pattern was also found for fruit size, since plant species bearing small to medium-sized fruits prevailed in all forest plots (> 60% of all species, Fig. 4). Furthermore, species bearing small-fruits were significantly reduced in numbers from younger to older forest plots (rs=0.76, n=16, P=0.0005), whereas the percentage of plant species bearing medium-sized fruits was positively correlated with the age of forest plots (rs=0.67, n=16, P < 0.004). In four sites sampled the percentage of woody species bearing either small seeds or fruits was significantly reduced from younger to older forest plots, whereas in three sites the percentage of plant species bearing medium-sized seeds and fruits significantly increased from younger to older plots (Table 3, within-site comparisons). 3.3. Plant family composition of forest plots In the early successional forest plots within the six sites sampled (Cubata˜o, Santa Virgı´nia, Macae´, Morretes,

Iporanga and Intervales), plant species belonging to the Melastomataceae, Rubiaceae, Flacourtiaceae and Myrsinaceae comprised between 25 and 42% of all species dispersed by any vertebrate. Species belonging to these families were far more common in these early successional plots than in old-growth plots (G=8.92; d.f.=1; P=0.003). On the other hand, species belonging to medium-seeded families such as the Lauraceae and Myrtaceae comprised between 33 and 41% of all vertebrate-dispersed species in the old-growth forests occurring in four sites sampled (Cubata˜o, Santa Virgı´nia, Macae´ de Cima and Morretes). The species richness of these families was therefore significantly higher in oldgrowth than in early successional forests (G=22.9; d.f.=1; P < 0.001). 3.4. Ordination of forest plots The first two axes of the DCA explained 67.6% of the variance within the species data at the level of families (sum of all eigenvalues=0.142, l1=0.075, l2=0.017). The figure of 67.6% of the total variance explained by the first two DCA axes of an unconstrained ordination compares favourably with that of most other neotropical forest plant communities examined to date. The next two axes of the DCA explained only a further 2.9% of the variance in the species data (l3=0.004, l4=0.000), and were thus discarded. Most late-successional and old-growth plots clustered together on the second DCA axis irrespective of the location at which they were sampled (Fig. 5). This convergence according to forest age thus provides marked evidence that the frequency of occurrence of Melastomataceae, Rubiaceae, Flacourtiaceae, Myrsinaceae,

M. Tabarelli, C.A. Peres / Biological Conservation 106 (2002) 165–176

171

Fig. 4. Relationship between the age of 16 montane Atlantic forest plots and the percentage of woody plant species in different classes of fruit size. Table 3 Percentage of vertebrate-dispersed woody species in different seed and fruit size-classes in the 16 plots sampled Site

Age of forest plots (year)

Cubata˜o Intervales Iporanga Macae´ de Cima Morretes Santa Virgı´nia * a

15, old growth 5, 13, 18 5, 15, 50 30, old growth 13, 15, old growth 18, 40, old growth

Percentage of plant speciesa Seeds <0.6 mm

Seeds 0.6–1.5 cm

Fruits <0.6 cm

Fruits 0.6–1.5 cm

44, 28.5* 71.4, 52.6, 54.5 72.2, 60.6, 40.5* 46.6, 44.5 58.8, 54.5, 35.1* 65.1, 43.4, 24.6*

30.0, 39.6 14.3, 36.8, 34.1 16.6, 33.3, 43* 39.6, 44.5 23.5, 24.2, 38.8 30.2, 42.1, 59.25*

24.4, 11.9 44.4, 36.8, 27.9 33.3, 37.5, 25.9 27.2, 23.2 58.8, 37.5, 16.3* 45.2, 28.5, 8.7 *

42.8, 55.2 33.3, 52.6, 51.1 27.7, 37.5, 44.4 40, 50.9 23.5, 28.1, 54.1* 28.5, 46.7, 63.7*

G-test pairwise comparisons, P <0.05. Ordered from the youngest to the oldest forest plot.

Lauraceae and Myrtaceae species varied according to the age of forest plots, rather than their geographic region, proximity or elevation. In particular, this convergence primarily reflects the greater density of Lauraceae and Myrtaceae tree species in late-successional and oldgrowth plots as opposed to young second-growth which tended to be largely dominated by understory shrub and small trees belonging to the Melastomataceae, Rubiaceae, Myrsinaceae and Flacourtiaceae. One latesuccessional plot (Iporanga), however, was identified as an outlier somewhat resembling the floristic composition of a younger forest plot sampled in the same region.

4. Discussion 4.1. Shifts in species composition, seed dispersal mode and diaspore size The data analysed here were gathered from several studies conducted on the basis of different methods and

sampling effort, and within plots of different size. It is possible that these differences may have introduced some bias in our results, mainly in the between-site comparisons which should be interpreted with caution. However, it is unlikely that differences in sampling methods and plot size would significantly influence the patterns uncovered as these were the same regardless of whether we compare plots within or between sites. Our results therefore support the notion that there are predictable shifts in species composition and seed dispersal strategies throughout the regeneration process of Atlantic montane forest. More specifically, increasingly older plots were associated with both a greater proportion of woody plant species relying on animal dispersal agents, and of vertebrate-dispersed species bearing medium-sized seeds and fruits. In contrast, early successional plots were associated with a greater proportion of species dispersed by abiotic means and of vertebrate dispersed-species bearing small seeds and fruits. Small-fruit species represented between 24.4 and 58.8% of the species in early

172

M. Tabarelli, C.A. Peres / Biological Conservation 106 (2002) 165–176

species sampled, respectively. At the Santa Virgı´nia Reserve this family represented between 8.9 and 27% of the species sampled in a 18-year-old and an old-growth forest, respectively. We therefore hypothesise that shifts in the relative importance of dispersal strategies (vertebrate vs. abiotic) during the regeneration process of Atlantic montane forest, as well as in diaspore size (small vs. mediumsized seeds and fruits) are related to the balance between early successional and shade-tolerant species associated with specific life forms (large shrubs vs. trees) and the frequency of species within particular plant families (Melastomataceae, Rubiaceae, Flacourtiaceae and Myrsinaceae vs. Lauraceae and Myrtaceae). 4.2. Implications for forest regeneration Fig. 5. Detrended correspondence analysis ordination showing the location of each forest plot along DCA axes I and II in relation to the species-richness of all tree families occurring at each plot. Forest plots younger and older than 40 years are indicated by open and solid circles, respectively. Site codes indicate the six sites (Fig. 1; Table 1) and the number of years elapsed since the last historical clear-cut (in parentheses).

secondary forests (5–30-year-old) and between 8.7 and 23.3% of the species in the old-growth forests. In the Atlantic montane forest, the most species-rich plant families associated with small seeds and fruits are the Melastomataceae, Rubiaceae, Flacourtiaceae and Myrsinaceae (Reitz, 1965; Silva and Tabarelli, 2000). These families occurred primarily within early secondary forests. Most of these species are represented by large shrubs or small trees (DBH < 10 cm) and are usually classified as pioneer or early successional species (cf. Klein, 1980; Denslow et al., 1990; Baider et al., 2001). These families typically produce sugar-rich berries or drupes smaller than 5 mm in diameter and the seeds of which are primarily dispersed by small frugivores (Moermond and Denslow, 1985; Dalling and Wirth, 1998). In contrast, species belonging to the Myrtaceae and Lauraceae comprised a large proportion of the flora bearing medium-sized seeds and fruits in the survey sites, including 33 and 41% of all vertebrate-dispersed species in the four old-growth plots. The most speciesrich genera in the forests surveyed—Eugenia, Myrcia, Marliera, Ocotea, Nectandra and Cryptocarya—are all shade-tolerant trees, even in early secondary forests (Vattimo, 1979; Tabarelli, 1997). However, the greatest levels of species richness for these genera are found within old-growth forests (Klein, 1980; Silva and Leita˜o Filho, 1982; Pessoa et al., 1997; Tabarelli and Mantovani, 1999a). For example, in the Morretes site we recorded one, three, and 16 species of Myrtaceae in increasingly older forests (13, 25-year-old and oldgrowth forest) accounting for 4, 7, and 21% of the

In the Atlantic forest and elsewhere in the neotropics, fruits of Melastomataceae, Rubiaceae, Flacourtiaceae and Myrsinaceae are usually consumed by small passerines such as manakins and tangaras (Pipra, Tachyphonus, Thraupis; Moermond and Denslow, 1985; Stiles and Rosselli, 1993; Sick, 1997, Loiselle and Blake, 1999; Develey and Peres, 2000). For example, at least seven tanager species are known to eat fruits of Leandra barbinervis, Leandra levigata and Miconia rigidiuscula (Melastomataceae species) in southern Atlantic montane forest (Rodrigues, 1995). In contrast, fleshy fruits of Myrtaceae and Lauraceae trees in the montane Atlantic forest are usually consumed by howler monkeys (Alouatta fusca) and woolly spider monkeys (Brachyteles arachnoides), cotingas (Procnias, Tijuca), toucans (Ramphastos), aracaris (Pteroglossus, Baillonius) and cracids (Penolope, Pipile; Petroni, 1993; Galetti, 1995; Galetti and Pizo, 1996; Ho¨fling and Camargo, 1996; Sick, 1997). For example, woolly spider monkeys, the largest neotropical primates, have been recorded feeding on fruits of at least 14 species of Myrtaceae in a montane forest reserve of southeast Brazil, and seed passage over long retention times through their digestive tract effectively increased germination rates of at least four Myrtaceae species (Moraes, 1992). Studies in other neotropical forests also suggest that seeds of Myrtaceae and Lauraceae trees are primarily dispersed by medium to large-bodied arboreal frugivores (i.e. primates and large-gaped birds). In French Guiana and Surinam, seeds of Eugenia and Camponesia trees, two species-rich Myrtaceae genera, are primarily dispersed by tamarins, capuchins, howlers and spider monkeys (van Roosmalen, 1985). Likewise, Myrtaceae seeds are frequently ingested by primates ranging in size from tamarins to woolly monkeys throughout lowland Amazonia (Terborgh, 1983; Peres, 1993, 1994). In Mesoamerican forests, quetzals, toucanets and bellbirds are the main dispersers of some 40 tree species of Lauraceae (Moermond and Denslow, 1985;

M. Tabarelli, C.A. Peres / Biological Conservation 106 (2002) 165–176

Wheelwright, 1985, 1986; Guindon, 1996; Wenny and Levey, 1998). In this study we did not quantify rates of seed removal for the Myrtaceae and Lauraceae by different animal species, or determine how seed dispersal might affect plant recruitment at both local and regional scale. However, if we accept that (1) seeds of Myrtaceae and Lauraceae trees are primarily dispersed by mid-sized to large frugivores and (2) some of these species could face severe recruitment bottlenecks in the absence of seed dispersal services (Wenny and Levey, 1998; Silva and Tabarelli, 2000), then it becomes imperative to examine the fabric of plant–animal interactions mediating the regeneration ecology of the Brazilian Atlantic forest since these two families contribute with a large proportion of the tree species richness throughout this region (Klein, 1980; Gentry, 1982; Mori et al., 1983; Leita˜o Filho, 1992; Tabarelli, 1997). The remaining Brazilian Atlantic forest cover is largely represented by small forest fragments surrounded by a hostile matrix of crops and pastures (Brown and Brown, 1992; Coimbra-Filho and Caˆmara, 1996; Ranta et al., 1998; Tabarelli et al., 1999). For example, the remaining forest cover in the state of Pernambuco consists of over 20,000 fragments, 48% of which are smaller than 10 ha and 85% are smaller than 50 ha (Ranta et al., 1998). Likewise, over 99% of the 100,000 forest fragments in the State of Rio de Janeiro are smaller than 500 ha (Gascon et al., 2000). Indeed several regions of the Atlantic forest are best described as huge archipelagos of small forest remnants in contrast to the sites analysed in this study, which contain the last large and continuous tracts of montane Atlantic forest. In these archipelagos, regenerating and mature forest remnants are vulnerable to continuing disturbance since most of them occur within private properties (Viana et al., 1997) and are subjected to shifting agriculture, selective logging and wildfires that have maintained a successional mosaic, particularly in the last 300 years (AndradeLima, 1970; Brown and Brown, 1992; Coimbra-Filho and Caˆmara, 1996; Tabarelli and Mantovani, 1999a). This means that much of the Atlantic forest regeneration is expected to occur after remnants are partially or totally converted into agricultural land and subsequently abandoned to revert to secondary forest. Unfortunately, local extinction of large vertebrates by habitat fragmentation, habitat loss and overhunting is an ongoing process over much of the remaining Atlantic forest (Almeida et al., 1995; Galetti et al., 1997; Chiarello, 1999; Cullen et al., 2000). In many cases, local extinction of large-bodied vertebrates are a result of both habitat fragmentation and subsistence hunting because small fragments accommodate small populations, are more accessible to hunters, and cannot be easily recolonised from surrounding source populations exposed to low levels of hunting pressure (Robinson,

173

1996; Silva and Tabarelli, 2000; Peres, 2001). Most of the Atlantic forest vertebrate species that are vulnerable to local extinction within fragments are large-bodied or relatively specialised frugivores (e.g. primates, cracids, cotingas, and toucans), whose seed dispersal services are often associated with a large proportion of medium to large-seeded plant species (e.g. Myrtaceae, Lauraceae). The composition of the seed disperser fauna appear to be one of the major filters for the seedling recruitment in successional forest landscapes since effective dispersal agents can greatly enhance successful colonisation of new secondary forest patches (Corlett and Turner, 1997; Duncan and Chapman, 1999). Previous studies in the old colonisation frontier of eastern Brazilian Amazonia have demonstrated that forest regeneration following cattle ranching and shifting agriculture can be restricted by the low availability of forest tree seeds (Uhl, 1987; Vieira et al., 1996). In the Paragominas region, trees with the highest establishment probability in abandoned pastures are those bearing seeds smaller than 0.18 g which are carried by small birds and bats (Nepstad et al., 1996). In a comparison between rare and abundant tree species recorded in a successional Amazonian forest mosaic, Vieira et al. (1996) found that trees classified as rare usually had seeds larger than 10 g. Similar dispersal-limitation scenarios have been proposed for other fragmented tropical forest regions where assemblages of large frugivorous vertebrates in small fragments appear to represent nested subsets of the regional species pool still retained by large fragments or remaining expanses of continuous forest (e.g. Kattan and Alvarez-Lo´pes, 1996; Restrepo et al., 1997; Warbuton, 1997; Chapman and Oderdonk, 1998). In light of the present state of most the Atlantic forest, our results should prompt additional studies on the extent to which: (1) seedling recruitment in the Myrtaceae and Lauraceae species is limited by a lack of effective seed dispersal services at both local and regional scales; (2) forest regeneration relies on seed delivery from adjacent forest patches and mediated by medium to large vertebrates; and (3) the increasingly fragmented Brazilian Atlantic forest faces widespread dispersallimitation scenarios. As a working hypothesis to be tested in further studies we conjecture that the species richness of woody plants in old secondary forests could eventually be reduced by as much as one half because of seed delivery and plant recruitment failures in Myrtaceae and Lauraceae species resulting from the local extirpation of medium to large vertebrates in highly fragmented portions of the Brazilian Atlantic forest.

Acknowledgements We thank the State of Pernambuco Research Council (Fundac¸a˜o de Amparo a` Cieˆncia e Tecnologia, FACEPE)

174

M. Tabarelli, C.A. Peres / Biological Conservation 106 (2002) 165–176

for funding this study. C. Peres’ work in neotropical forests was funded by the Center for Applied Biodiversity Science of Conservation International and the Josephine Bay and Michael Paul Foundation. We also thank M. Schwartz and W. F. Laurance for constructive criticisms on the manuscript.

References Almeida, R.T., Pimentel, D.S., Silva, E.M.C., 1995. The red-handed howling monkey in the State of Pernambuco, north-east Brazil. Neotropical Primates 3, 174–176. Andrade-Lima, D., 1970. Recursos vegetais de Pernambuco. Boletim do IPA 41, 1–32. Baider, C.L., Tabarelli, M., Mantovani, W., 2001. The soil seed bank during Atlantic forest regeneration in southeast Brazil. Revista Brasileira de Biologia 61, 35–44. Barroso, G.M., Morim, M.P., Peixoto, A.L., Ichaso, C.L.F., 1999. Frutos e Sementes: Morfologia Aplicada a` Sistema´tica de Dicotiledoˆneas. Universidade Federal de Vic¸osa, Vic¸osa. Brown Jr., K.S., Brown, G.G., 1992. Habitat alteration and species loss in Brazilian forests. In: Whitmore, T.C., Sayer, J.A. (Eds.), Tropical Deforestation and Species Extinction. Chapman & Hall, London, pp. 129–142. Brown, S., Lugo, A.E., 1990. Tropical secondary forests. Journal of Tropical Ecology 6, 1–32. Chapman, C.A., Onderdonk, D.A., 1998. Forest without primates: primate/plant codependency. American Journal of Primatology 45, 127–141. Charles-Dominique, P., 1993. Speciation and coevolution: an interpretation of frugivory phenomena. Vegetatio 107/108, 75–85. Chiarello, A.G., 1999. Effects of fragmentation of the Atlantic forest mammal communities in south-eastern Brazil. Biological Conservation 89, 71–82. Coimbra-Filho, A.F., Caˆmara, I.G., 1996. Os Limites Originais do Bioma Mata Atlaˆntica na Regia˜o Nordeste do Brasil. Fundac¸a˜o Brasileira de Conservac¸a˜o da Natureza, Rio de Janeiro. Corlett, R.T., 1998. Frugivory and seed dispersal by vertebrates in the Oriental (Indomalayan) Region. Biological Review 73, 413–448. Corlett, R.T., Turner, I.M., 1997. Long-term survival in tropical forest remnants in Singapore and Hong Kong. In: Laurance, W.F., Bierregaard Jr., R.O. (Eds.), Tropical Forest Remnants: Ecology, Management, and Conservation of Fragmented Communities. The University of Chicago Press, Chicago, pp. 333–346. Cox, P.A., Elmqvist, T., Pierson, E.D., Rainey, W.E., 1991. Flying foxes as strong interactors in South Pacific Island ecosystems: a conservation hypothesis. Conservation Biology 5, 448–454. Cullen Jr., L., Bodmer, R.E., Padua, C.V., 2000. Effects of hunting in habitat fragments of the Atlantic forests, Brazil. Biological Conservation 95, 149–156. Dalling, J.W., Wirth, R., 1998. Dispersal of Miconia argenta seeds by the leaf-cutting ant Atta colombica. Journal of Tropical Ecology 14, 705–710. Denslow, J.S., Schultz, J.C., Vitousek, P.M., Strain, B.R., 1990. Growth responses of tropical shrubs to treefall gap environments. Ecology 71, 165–179. Develey, P.F., Peres, C.A., 2000. Resource seasonality and the structure of mixed-species bird flocks in a coastal Atlantic forest of southeastern Brazil. Journal of Tropical Ecology 16, 33–53. Duncan, R.S., Chapman, C.A., 1999. Seed dispersal and potential forest succession in abandoned agriculture in tropical Africa. Ecological Applications 9, 998–1008. Galetti, M., 1995. Os frugı´voros de Santa Genebra. In: Morellato, P.C., Leita˜o Filho, H.F. (Eds.), Ecologia e Preservac¸a˜o de uma

Floresta Tropical Urbana, Reserva de Santa Genebra. Editora da UNICAMP, Campinas, pp. 66–68. Galetti, M., Pizo, M.A., 1996. Fruit eating by birds in a forest fragment in southeastern Brazil. Ararajuba 4, 71–79. Galetti, M., Martuscelli, P., Olmos, F., Aleixo, A., 1997. Ecology and conservation of the Jacutinga Pipile jacutinga in the Atlantic forest of Brazil. Biological Conservation 82, 31–39. Gascon, C., Williamson, B., Fonseca, G.A.B., 2000. Receding forest edges and vanishing reserves. Science 288, 1356–1358. Gentry, A.H., 1982. Patterns of neotropical plant species diversity. Evolution Biology 15, 1–84. Gentry, A.H., 1995. Diversity and floristic composition of neotropical dry forests. In: Bullock, S.H., Mooney, H.A., Medina, E. (Eds.), Seasonal Dry Tropical Forests. Cambridge University Press, London, pp. 146–190. Guapyassu, M.S., 1994. Caracterizac¸a˜o fitossociolo´gica de treˆs fases sucessionais de uma floresta ombro´fila densa submontana, Morretes-Parana´. MSc thesis, Setor Cieˆncias Agra´rias, Universidade Federal do Parana´, Parana´, Brazil. Guaricata, M.R., Pinard, M.A., 1998. Ecological knowledge of regeneration from seed in neotropical forest trees: implications for natural forest management. Forestry Ecology and Management 11, 87–99. Guedes-Bruni, R.R., Pessoa, S.V.A., Kurtz, B.C., 1997. Florı´stica e estrutura do componente arbustivo-arbo´reo de um trecho preservado de floresta montana na Reserva Ecolo´gica de Macae´ de Cima. In: Lima, H.C., Guedes-Bruni, R.R. (Eds.), Serra de Macae´ de Cima: Diversidade Florı´stica e Conservac¸a˜o em Mata Atlaˆntica. Jardim Botaˆnico do Rio de Janeiro, Rio de Janeiro, pp. 127–146. Guindon, F.C., 1996. The importance of forest fragments to the maintenance of regional biodiversity in Costa Rica. In: Schellas, J., Greenberg, R. (Eds.), Forest Patches in Tropical Landscapes. Island Press, London, pp. 163–186. Hardwick, K., Healey, J., Elliott, S., Garwood, N., Anusarnsunthorn, V., 1997. Understanding and assisting natural regeneration processes in degraded seasonal evergreen forests in northern Thailand. Forestry Ecology and Management 99, 203–214. Hill, M.O., Gauch, H.G., 1980. Detrended correspondence analysis, an improved ordination technique. Vegetatio 42, 47–58. Ho¨fling, E., Camargo, H.F.A., 1996. Aves do Campus. Edusp, Sa˜o Paulo. Howe, H.F., Smallwood, J., 1982. Ecology of seed dispersal. Annual Review of Ecology and Systematics 13, 201–228. Kattan, G., Alvarez-Lo´pez, H., 1996. Preservation and management of biodiversity in fragmented landscapes in the Colombian Andes. In: Schellas, J., Greenberg, R. (Eds.), Forest Patches in Tropical Landscapes. Island Press, London, pp. 3–18. Klein, R.M., 1980. Ecologia da flora e vegetac¸a˜o do Vale do Itajaı´. Sellowia 32, 165–389. Leita˜o Filho, H.F., 1992. A flora arbo´rea da Serra do Japi. In: Morellato, L.P.C. (Ed.), Histo´ria natural da Serra do Japi: Ecologia e Preservac¸a˜o de uma A´rea Florestal no Sudeste do Brasil. Editora da UNICAMP e FAPESP, Sa˜o Paulo, pp. 40–43. Leita˜o Filho, H.F. (Ed.), 1993. Ecologia da Mata Atlaˆntica em Cubata˜. Editora da UNESP e Editora da UNICAMP, Sa˜o Paulo. Lima, A.R., Capobianco, J.P.R., 1997. Mata Atlaˆntica: avanc¸os legais e institucionais para sua conservac¸a˜o. Documentos do ISA No. 4. Instituto Socioambiental, Sa˜o Paulo. Lima, H.C., Guedes-Bruni, R.R., 1997. Serra de Macae´ de Cima: Diversidade Florı´stica e Conservac¸a˜o em Mata Atlaˆntica. Jardim Botaˆnico do Rio de Janeiro, Rio de Janeiro. Loiselle, B.A., Blake, J.G., 1999. Dispersal of melastome seeds by fruit-eating birds of tropical forest understory. Ecology 80, 330–336. Lorenzi, H., 1998. A´rvores Brasileiras: Manual de Identificac¸a˜o e Cultivo de Plantas Arbo´reas Nativas do Brasil. Vol. 1 e 2. Editora Plantarum, Nova Odessa.

M. Tabarelli, C.A. Peres / Biological Conservation 106 (2002) 165–176 Melo, M.R.F., Mantovani, W., 1994. Composic¸a˜o florı´stica e estrutura fitossociolo´gica da Mata Atlaˆntica de encosta na Ilha do Cardoso (Canane´ia, Brasil). Boletim do Instituto de Botaˆnica 9, 107–158. Moermond, T.C., Denslow, J.S., 1985. Neotropical avian frugivores: patterns of behavior, morphology, and nutrition, with consequences for fruit selection. In: Buckley, P.A., Foster, M.S., Morton, E.S., Ridgely, R.S., Buckley, F.G. (Eds.), Neotropical Ornithology. Ornithological Monographs No 36. The American Ornithologists’ Union, Washington, DC, pp. 865–896. Moraes, P.L.R., 1992. Dispersa˜o de sementes pelo mono-carvoeiro (Brachyteles arachnoides E. Geoffroy, 1806) no Parque Estadual de Carlos Botelho. In: II Congresso Nacional sobre Esseˆncias Nativas, Sa˜o Paulo, Anais 4, pp. 1199–1205. Mori, S.A., Boom, B.M., Carvalino, A.M., Santos, T.S., 1983. Ecological importance of Myrtaceae in a eastern Brazilian wet forest. Biotropica 15, 68–70. Myers, N., Mittermeier, R.A., Mittermeier, C.G., Fonseca, G.A.B., Kent, J., 2000. Biodiversity hotspots for conservation priorities. Nature 403, 845–853. Nascimento, F.A., 1994. A sucessa˜o secunda´ria inicial na mata atlaˆntica, sobre a Serra de Paranapiacaba, Ribeira˜o Grande, SP. MSc thesis, Instituto de Biocieˆncias, Univerisdade de Sa˜o Paulo, Sa˜o Paulo, Brazil. Nepstad, D., Uhl, C., Pereira, C.A., Silva, J.M.C., 1996. A comparative study of tree establishment in abandoned pasture and mature forest of eastern Amazonia. Oikos 76, 25–39. Oliveira, R.R., Sampaio, P.D., Silva, R.F., Toffoli, D.G., 1994. Roc¸a caic¸ara, um sistema ‘‘primitivo’’ auto-sustenta´vel. Cieˆncia Hoje 18, 44–51. Peres, C.A., 1993. Diet and feeding ecology of saddle-back and moustached tamarins in an Amazonian terra firme forest. Journal of Zoology 230, 567–592. Peres, C.A., 1994. Diet and feeding ecology of gray woolly monkeys (Lagothrix lagotricha cana) in central Amazonia: comparisons with other atelines. International Journal of Primatology 15, 333–372. Peres, C.A., 2001. Synergistic effects of subsistence hunting and habitat fragmentation on Amazonian forest vertebrates. Conservation Biology 15 (6), 1490–1505. Peres, C.A., van Roosmalen, M.G.M., 2002. Patterns of primate frugivory in Amazonia and the Guianan shield: implications to the demography of large-seeded plants in overhunted forests. In: Levey, D., Galetti, M., Silva, W. (Eds.), Frugivory and Seed Dispersal: Ecological, Evolutionary and Conservation Issues. CABI Publishing, Oxford (in press). Pessoa, S.V.A., Guedes-Bruni, R.R., Bruno, C.K., 1997. Composic¸a˜o florı´stica e estrutura do componente arbustivo-arbo´reo de um trecho secunda´rio de floresta montana na Reserva Ecolo´gica de Macae´ de Cima. In: Lima, H.C., Guedes-Bruni, R.R. (Eds.), Serra de Macae´ de Cima: Diversidade Florı´stica e Conservac¸a˜o em Mata Atlaˆntica. Jardim Botaˆnico do Rio de Janeiro, Rio de Janeiro, pp. 147–168. Petroni, L.M., 1993. Aspectos da ecologia e comportamento do mono-carvoeiro, Brachyteles arachnoides (E. Geoffroy, 1806), (Cebidae-Primates), na Fazenda Intervales, Serra de Paranabiacaba, SP. MSc thesis, Pontifı´cia Universidade Cato´lica, Porto Alegre, Brazil. Radambrasil, 1983. Projeto Radambrasil: Levantamento de Recursos Naturais. IBGE, Rio de Janeiro. Ranta, P., Blom, T., Niemela, J., Joensuu, E., Siitonen, M., 1998. The fragmented Atlantic rain forest of Brazil: size, shape and distribution of forest fragments. Biodiversity and Conservation 7, 385–403. Reitz, R., 1965. Flora Ilustrada Catarinense. Herba´rio Barbosa Rodrigues, Itajaı´, SC. Restrepo, C., Renjifo, L.M., Marples, P., 1997. Frugivorous birds in fragmented neotropical montane forests: landscape pattern and body mass distribution. In: Laurance, W.F., Bierregaard Jr., R.O.

175

(Eds.), Tropical Forest Remnants: Ecology, Management, and Conservation of Fragmented Communities. The University of Chicago Press, Chicago, pp. 171–189. Richards, P.W., 1996. The Tropical Rain Forest, 2nd Edition. Cambridge University Press, Cambridge. Robinson, J.G., 1996. Hunting wildlife in forest patches: an ephemeral resource. In: Schellas, J., Greenberg, R. (Eds.), Forest Patches in Tropical Landscapes. Island Press, London, pp. 111–130. Rodrigues, M., 1995. Spatial distribution and food utilization among tanagers in southeastern Brazil (Passeriformes: Emberizidae). Ararajuba 3, 27–32. Sick, H., 1997. Ornitologia Brasileira. Editora Nova Fronteira, Rio de Janeiro. Siegel, S., 1979. Estatı´stica Na˜o-Parame´trica. Editora McGraw-Hill do Brasil Ltda, Sa˜o Paulo. Silva, A.F., Leita˜o Filho, H.F., 1982. Composic¸a˜o florı´stica e estrutura de um trecho de Mata Atlaˆntica de encosta no municı´pio de Ubatuba (Sa˜o Paulo, Brasil). Revista Brasileira de Botaˆnica 5, 43–52. Silva, J.M.C., Tabarelli, M., 2000. Tree species impoverishment and the future flora of the Atlantic forest of northeast Brazil. Nature 404, 72–74. Sokal, R.R., Rohlf, F.G., 1995. Biometry. W.H. Freeman and Company, New York. Stiles, F.G., Rosselli, L., 1993. Consumption of fruits of the Melastomataceae: how diffuse is coevolution. Vegetatio 107/108, 57–73. Tabarelli, M., 1997. A regenerac¸a˜o da floresta Atlaˆntica montana. PhD thesis, Instituto de Biocieˆncias, Universidade de Sa˜o Paulo, Sa˜o Paulo, Brazil. Tabarelli, M., Mantovani, W., 1999a. A regenerac¸a˜o de uma floresta tropical montana apo´s corte e queima (Sa˜o Paulo-Brasil). Revista Brasileira de Biologia 59, 239–250. Tabarelli, M., Mantovani, W., 1999b. A riqueza de espe´cies arbo´reas na floresta atlaˆntica de encosta de Sa˜o Paulo (Brasil). Revista Brasileira de Botaˆnica 22, 217–223. Tabarelli, M., Mantovani, W., Peres, C.A., 1999. Effects of habitat fragmentation on plant guild structure in the montane Atlantic forest of southeastern Brazil. Biological Conservation 91, 119–127. Terborgh, J., 1983. Five New World Primates: a Study in Comparative Ecology. Princeton University Press, Princeton. ter Braak, C.J.F., 1987–1992. CANOCO—a FORTRAN program for canonical community ordination. Microcomputer Power, Ithaca, New York. The´baud, C., Strasberg, D., 1997. Plant dispersal in fragmented landscapes: a field study of woody colonization in rainforest remnants of the Mascarene Archipelago. In: Laurance, W.F., Bierregaard Jr., R.O. (Eds.), Tropical Forest Remnants: Ecology, Management, and Conservation of Fragmented Communities. The University of Chicago Press, Chicago, pp. 321–332. Torezan, J.M.D. Estudo da sucessa˜o secunda´ria, na floresta ombro´fila densa sub-montana, em a´reas anteriormente cultivadas pelo sistema de ‘‘coivara’’, em Iporanga—SP. MSc thesis, Setor de Cieˆncias Agra´rias, Universidade Federal do Parana´, Parana´, Brazil. Uhl, C., 1987. Factors controlling succession following slash-and-burn agriculture in Amazonia. Journal of Ecology 75, 377–407. van der Pijl, L., 1982. Principles of Dispersal in Higher Plants, 3rd Edition. Springer-Verlag, Berlin. van Roosmalen, M.G.M., 1985. Fruits of the Guianan Flora. Institute of Systematic Botany, Utrecht University, Netherlands. Vattimo, I., 1979. Laura´ceas. Flora Ilustrada Catarinense, Herba´rio Barbosa Rodrigues, Itajaı´. Veloso, H.P., Rangel-Filho, A.L.R., Lima, J.C.A., 1991. Classificac¸a˜o da Vegetac¸a˜o Brasileira Adaptada a um Sistema Universal. IBGE, Rio de Janeiro. Viana, V.M., Tabanez, A.J., Batista, J.L., 1997. Dynamics and restoration of forest fragments in the Brazilian Atlantic moist forest. In: Laurance, W.F., Bierregaard Jr., R.O. (Eds.), Tropical Forest

176

M. Tabarelli, C.A. Peres / Biological Conservation 106 (2002) 165–176

Remnants: Ecology, Management, and Conservation of Fragmented Communities. The University of Chicago Press, Chicago, pp. 351–365. Vieira, I.C.G., Saloma˜o, R.P., Rosa, N.A., Nepstad, D.C., Roma, J.C., 1996. O renascimento da floresta no rastro da agricultura. Cieˆncia Hoje 20, 38–44. Warburton, N.H., 1997. Structure and conservation of forest avifauna in isolated rainforest remnants in tropical Australia. In: Laurance, W.F., Bierregaard Jr., R.O. (Eds.), Tropical Forest Remnants: Ecology, Management, and Conservation of Fragmented Communities. The University of Chicago Press, Chicago, pp. 106–190.

Wenny, D.G., Levey, D., 1998. Directed seed dispersal by bellbirds in a tropical cloud forest. Proceedings of the National Academy of Sciences 95, 6204–6207. Wheelwright, N.T., 1985. Competition for dispersers, and the timing of flowering and fruiting in a guild of tropical trees. Oikos 44, 465–477. Wheelwright, N.T., 1986. A seven-year of individual variation in fruit production in tropical bird-dispersed tree species in the family Lauraceae. In: Estrada, A., Fleming, T.H. (Eds.), Frugivores and Seed Dispersal. Dr W. Junk Publishers, Dordrecht, pp. 19–35. Wilkinson, L., 1996. SYSTAT, version 6.0. SPSS, Chicago, Illinois, USA.