Introduction

Fragmentation may result in rapid population growth in mammal species whose population sizes, in other more diverse communities, would have been controlled by competition or predation (Redford, 1992; Peres and Dolman 2000; Terborgh et al., 2001; Link et al., 2010). At least three hypotheses may explain these increased densities: (1) absence of predators, (2) ecological plasticity and (3) density compensation, in which the effects of isolation result in the elimination of some species (Redford, 1992; Glanz, 1996; González-Soíls et al., 2001; Terborgh et al., 2001). For example, in the Amazon, small and mediumsized primate species increase in abundance where larger species of the family Atelidae are preferentially hunted (Peres and Dolman, 2000). Also, greater densities in Alouatta and Sapajus species may be found in degraded communities in fragments and islands (Chiarello and Galetti 1994; González-Solís et al., 2001; Ludwig et al., 2005; Martins, 2005; Almeida-Silva et al., 2005; Link et al., 2010).

Islands may serve as natural experiments to examine the effects of isolation on mammal populations (Glanz, 1996; Terborgh et al., 2001). For example, abundance of Alouatta seniculus increased in the absence of predators as well as with simplification of forests on artificial islands (Terborgh et al., 2001). Similarly, black-and-gold howler monkey (Alouatta caraya) densities are greater on an island than the nearby shores of the Paraná River in Argentina (Rumiz, 1990; Zunino et al., 2001). However, islands in the Paraná River are natural and have primate predators, such as large cats (Aguiar et al., 2007; Ludwig et al., 2007). On the islands of the Paraná River, with rich alluvial soils, two important factors may contribute to high howler population densities: (1) greater productivity of high-quality resources, such as fruits and young leaves, and (2) year-round resource availability (Janzen, 1974; Rumiz, 1990; Zunino et al., 2001; Bravo and Sallenave, 2003).

Although A. caraya population studies have taken place in the system of islands of the Middle Paraná River (Rumiz, 1990; Zunino et al., 2001), few studies include other primates in this area (Brown and Zunino, 1994) and in others portions of this river. Given the importance of understanding local variation in primate abundances, the main goal of this study was to examine and compare densities of A. caraya, Sapajus nigritus and Sapajus cay on forested islands and along the adjacent margins of the Upper Paraná River in southern Brazil.

Methods

Study area

Primates were studied in the Upper Paraná River Protected Area in the southern Brazilian Atlantic Forest. Primate abundances were estimated on two islands and along the margins of the river near the city of Porto Rico in Paraná, and near the city of Taquarussu in Mato Grosso do Sul 22°43′60″S and 53°24′18″W; Fig. 1). The region is subtropical, with an average rainfall between 1,200–1,300 mm (Romagnolo and Souza, 2000). Human impact has degraded the forests, although today the area is in process of recovery (Campos and Souza, 2002). The study area comprises two main forest types: alluvial, or flooded forests (várzeas), that are found in the islands and flood plains on the Mato Grosso do Sul side of the river; and submontane (higher than 250 m above sea level) on the Paraná side of the river. Alluvial forests flood annually and have low tree diversity and are dominated by pioneer species (Romagnolo and Souza, 2000; Campos and Souza, 2002). The canopy is open and low (10–15 m, emergent to 25 m). Productivity is high due to the rich alluvial soils and so high-quality forage (leaves) is abundant and relatively constant for folivorous howler monkeys (Rumiz, 1990). Cecropia pachystachya is the dominant tree and the main food source for howler monkeys here (Ludwig et al., 2008). Submontane forests are the most diverse and mature forests in the region in areas that do not flood (M. C. Souza, unpublished). However, productivity of resources used by folivorous primates is relatively lower and more seasonal (Rumiz, 1990). The canopy is more closed with trees of 15–30 m. Today, the submontane forest is nearly extinct in this region with only a few fragments remaining (Campos and Souza, 1997).

Primates were counted on two islands and both margins of the river (Fig. 1). Mutum Island is the largest island (1,050 ha), near the river center (1,000 m from Paraná, 700 m from Mato Grosso do Sul) and comprises a mosaic of alluvial forests, swamps, and patches of more preserved forests, and is dominated by the pioneer tree C. pachystachya. Carioca Island is smaller (360 ha), 600 m from Paraná, 800 m from Mutum Island, and comprises a mosaic of degraded alluvial forests with few trees and discontinuous canopy with many vines. Humans and feral domestic pigs (Sus scrofa) have impacted the island. The shores of Mato Grosso do Sul comprise alluvial forests that are naturally narrow corridors above the dikes formed by the river. Today, they are fragmented and poorly preserved due to an increasing pressure from agriculture, cattle, fire, and human settlements. Farther from the dikes, the forests are rapidly being replaced by pasture and marshes. Finally, the Paraná shore has submontane forests and, despite some anthropic influence, has continuous, wider stretches that retain most of the original forest structure (M. C. Souza, unpublished). Abundant palms (Acrocomia aculleata and Syagrus romanzoffianum), and agriculture, including corn and cassava, are only on the Paraná side of the river (L. M. Aguiar, personal observation).

Figure 1.

Location of the study area on the Upper Paraná River, southern Brazil. X marks the transect locations on the river shore; MI: Mutum Island; CI: Carioca Island. Porto Rico: nearby city.

Results

A total of 397 sightings were recorded during about 500 h of walking over 188 km (Table 1). Howler monkey density varied between 0.40–2.56 ind ha^-1, with the lowest densities on the Paraná side and on Carioca Island, and the greatest on the Mato Grosso do Sul side and Mutum Island. Capuchin monkey density varied between 0.31 ind ha^-1 for S. cay in Mato Grosso do Sul and 0.51 ind ha^-1 for S. nigritus in Paraná.

Table 1.

Sampling effort (distance walked in km), number of sightings, sighting rate (sightings km-1), average group (cluster) size, and density estimate (individuals ha-1) at four sampling locations on the Upper Paraná River, in southern Brazil.

Discussion

The flooded, and more conserved forests on Mutum Island are apparently the most favorable local habitats for A. caraya such that their density is more than five times that of the other locations. Primate density is similar in the other locations, although capuchin density was greater in more mature and conserved forests (e.g., S. nigritus in Paraná) and greater than the density of sympatric howlers. Indeed, howler density was lower than that of the capuchins in mature forests of Paraná and in the most disturbed habitat of Carioca Island. Howler abundance, therefore, may be strongly influenced by productivity (such as flooded forests), while the more generalist capuchins do better in mature forests and more conserved habitats. We suggest that the flooded habitat in Mutum Island is important to maintain the extremely abundant C. pachystachya as a very important resource that allows greater population sizes in A. caraya. In contrast, while often flooded, Carioca Island apparently cannot support similar population sizes there due to environmental degradation. A similar relationship of abundance, productivity and degradation has been found in the Middle Paraná River (Rumiz, 1990; Brown and Zunino, 1994; Zunino et al., 2001).

Other factors may also influence densities in this study area: 1) low predation and hunting pressure, 2) density compensation (Redford, 1992; Glanz, 1996; González-Solís et al., 2001; Terborgh et al., 2001) and 3) crowding (Lovejoy et al., 1986). First, hunting pressure by humans is low and natural primate predators are less abundant in this region, probably due to the presence of humans. While predation is an important component of primate life histories, predation as a constraint on primate population size has not been clearly demonstrated in most species (Isbell, 1994). Density compensation may also allow greater abundance. Since the local primate community is depauperate, with never more than two co-occurring species in any one place, it is possible that reduced competition for resources and isolation due to fragmentation combine to favor increased density. Also, other potential competitors for howlers, such as capuchins and coatis are absent from the islands (Aguiar et al., 2007). This distribution raises the question of why there are not three primate species, and coatis, at any given site, since they are all locally available?. Perhaps the river is an effective barrier for dispersal of capuchins and coatis, or flooded forests may not provide the additional resources (prey, fruit) that these omnivorous species require (Haugaasen and Peres, 2005; Aguiar et al., 2011).

The crowding effect (Lovejoy et al., 1986) may also be important, since these primates are opportunists and generalists, and may have overlapping ranges. They may coexist at relatively high densities in environmentally degraded areas due to dispersal limitation, such as on the islands. This is an important possibility, because it suggests that degraded areas may still provide resources to maintain large or dense populations. Usually, capuchins are more frugivorous and more resource generalists than howlers (Freese and Oppenheimer, 1981; Link et al., 2010). However, S. cay is not very abundant (in contrast to howlers), apparently because the flooded forest is not ideal habitat for capuchin species (Peres, 1989; Haugaasen and Peres, 2005). Riparian forest along the river in Mato Grosso do Sul is a narrow corridor and with a relatively small area for foraging. In contrast, S. nigritus is more abundant on the Paraná side of the river than is typical (Ludwig et al., 2005; Martins, 2005; Almeida-Silva et al., 2005). On the Paraná side, the riparian forest corridor is wider, with more mature and better-preserved forest including many palm trees where they forage, as well has having cultivated crops that they may also consume (Ludwig et al., 2005). These complex interactions require autecological studies that examine the importance of each resource to better understand how these resources influence abundance patterns.

Therefore, here, the abundance patterns of these three primate species suggest that forest structure and conservation status may be the main influences regulating population size. Greater abundance may be explained by both, immigration and reproduction (Begon et al., 2006). Since howlers can swim (Aguiar et al., 2007) and then find abundant resources, their populations will grow as long as resources are not limiting. If riparian areas are well protected, they can maintain populations of primates and other animals. We suggest that the interactions of density compensation, crowding, low hunting and predation pressures, and the extremely abundant Cecropia (for howlers) probably allow the unusually high densities. Although humans have fragmented the habitat, primates are still relatively abundant, which suggests that these are important populations for conservation, in contrast with the lower abundances of these species elsewhere (Ludwig et al., 2005; Moura, 2007).

Thus, populations along the Paraná River are potentially important sources for regional maintenance of these species. We recommend further study to specifically tests the hypotheses presented here and thereby assure conservation of these primates.