Introduction

The geographic ranges of Amazonian primates have traditionally been defined on the basis of the distribution of major river systems, which play a fundamentally important role in the zoogeography of most genera (Ayres and Clutton-Brock 1992). This tradition was reinforced by the fact that, until very recently, the vast majority of recorded localities were clustered along the banks of major rivers (see for example, Hershkovitz 1977, 1990). It has also reinforced the implicit assumption that primate populations are distributed more or less uniformly within the limits defined by these rivers.

In recent years, the more systematic exploration of Amazonian interfluvia, facilitated in many cases by the construction of highways, has updated our knowledge not only of the zoogeography, but also of the diversity of many platyrrhine genera, most notably Mico (see Van Roosmalen et al. 2000) and Callicebus (see Van Roosmalen et al. 2002; Wallace et al. 2006). Detailed surveys have also revealed local variations in distribution and abundance, related primarily to ecological factors such as habitat characteristics, in particular differences between flooded or floodplain and terra firma forests, and interspecific competition (Ferrari and Lopes 1990, 1996; Peres 1993,1997a; Iwanaga and Ferrari 2002; Ferrari 2004).

The Tocantins and Xingu rivers are the principal barriers to the dispersal of primates in southeastern Amazonia, where they delimit the distribution of a number of platyrrhine taxa, including four genera (Ferrari and Lopes 1996). The Tocantins forms the easternmost limit of the distribution of Amazonian titis, and the red-bellied titi, Callicebus moloch, is known from a total of ten localities east of the Xingu (Mascarenhas and Puorto 1988; Ferrari and Lopes 1990; Hershkovitz 1990). However, detailed surveys at three sites in the Xingu-Tocantins interfluvium, presented here, indicate that C. moloch is either extremely rare in or absent from large tracts of forest within this area. The patchy distribution of the species appears to be a natural phenomenon, rather than a result of habitat disturbance or hunting pressure, but the factors involved remain unclear.

Methods

Primate populations were surveyed at three sites in the Tocantins-Xingu interfluvium (Fig. 1, Table 1) in 1996 and 1997. All three sites present relatively large tracts of primary Amazonian terra firma forest, but also patches of both secondary forest and distinct ecosystems such as inundated, or igapó forest (especially at the Ferreira Penna Scientific Station [ECFPn], in the Caxiuanã National Forest). Detailed descriptions of sites 1, 2 and 3 are given, respectively, in Lisboa (1997, 2001), Bobadilla (1998) and Emidio-Silva (1998). The principal difference between the sites is related to their location within the interfluvium, which can be divided into a lowland floodplain to the north, and the Brazilian Shield to the south, with associated differences in forest structure and composition (Brazil, MME-DNPM Projeto RADAM 1974; Ferrari and Lopes 1990, 1996; Lisboa et al. 1997). Site 1 represents the lowland floodplain, whereas sites 2 and 3 are located on the Brazilian Shield.

As in previous studies in southeastern Amazonia (for example, Johns 1986; Ferrari and Lopes 1996; Lopes and Ferrari 2000; Ferrari et al. 2003), primate populations were surveyed using standard line transect methods (see Brockelman and Ali 1986). A straight-line trail system was established in the terra firma forest at each location, with a total length of 65, 15.5 and 9.5 km, respectively, at sites 1, 2 and 3. Trails were cleared causing a minimum of impact, swept clean of debris and marked with flagging at 100 m intervals. During surveys, in 1996 and 1997, trails were walked at an average speed of 1–1.5 km per hour and, at each encounter with a primate group, the species and composition of the group were recorded, in addition to other information not relevant here (see Bobadilla and Ferrari 2000).

Results

A total of 292 sightings of primates were recorded during the 812 km surveyed at the three study sites (Table 2), at an average rate of 3.6 sightings per 10 km of transect. This sighting rate compares favorably with those recorded at other sites in southeastern Amazonia (Lopes and Ferrari 2000; Ferrari et al. 2002, 2003). Four species, Alouatta belzebul, Cebus apella, Chiropotes satanas, and Saguinus niger, were recorded at all three sites, and provided the vast majority (287) of sightings. A fifth species, Saimiri sciureus, was not seen at site 1 during the present study, but it has been observed within the study area. The overall lack of records of Saimiri may have been at least partly due to its specific habitat preferences and ranging behavior (Terborgh 1983).

One other species, Mico argentatus, has a limited distribution in the Xingu-Tocantins interfluvium, where it is restricted to the lowland floodplain (Ferrari and Lopes 1990), and is thus absent from sites 2 and 3. The species occurs at site 1 (Ferrari and Lopes 1996), but is found exclusively in secondary forest habitats, where it may reach relatively high densities (Veracini 1997). It has never been observed within the present study area.

The total lack of records of Callicebus moloch from all three sites is less easily accounted for. Titis appear to be at least as abundant at other sites in this interfluvium (Mascarenhas and Puerto 1988; Ferrari and Lopes 1990) as they are west of the Xingu (Martins et al. 1988; Ferrari et al. 2003) or elsewhere in the Amazon basin (Peres 1997a; Ferrari et al. 2000).

Ferrari et al. (2003) recorded C. moloch in both continuous and fragmented forest east of the Rio Tapajós, normally at median densities, although the species was absent from some sites, possibly due to the effects of habitat fragmentation. Even if local conditions at the present study sites were to reduce the visibility of the species during surveys for some reason, it seems unlikely that they would also suppress the characteristic vocal duetting that is typical of all titi species (Emmons et al. 1998). Audible over distances of more than one kilometer, and normally performed on a daily basis, the duet is a sure sign of the presence of titis at most sites.

There is also little evidence of hunting pressure at any site, not least because titis are almost never targeted because of their small body size. The relative abundance of the larger species (Alouatta, Cebus and Chiropotes) at all three sites also indicates that primates suffer little pressure. The Parakanã do hunt primates, but very rarely, and only Alouatta and Cebus—coincidentally, the two species recorded most frequently at site 3.

Additional evidence confirms that C. moloch is absent from large areas of continuous forest at sites 1 and 3. Perhaps the most reliable evidence is that provided by experienced Parakanã hunters, who unanimously confirm the absence of Callicebus from site 3. However, as more than half of the Parakanã territory is still uninhabited (Emidio-Silva 1998), it is probably premature to exclude C. moloch from the whole of the reservation. Similarly, while the sum of the evidence from site 1 (ECFPn)— which now includes a number of longterm studies of primate ecology (Veracini 1997; Jardim and Oliveira 1997; Pina et al. 2001; Tavares and Ferrari 2001) and additional surveys (Martins et al. 2005)— leaves little doubt as to the absence of Callicebus, some local residents have reported its presence in areas neighboring the Caxiuanã National Forest. The results from both sites 2 and 3 also contradict considerably those of the rescue operation in the area of the Tucuruí reservoir (Mascarenhas and Puerto 1988), less than 50 km to the east (Figure 1), where Callicebus moloch was the third most frequently captured species, after Alouatta belzebul and Cebus apella.

Table 1.

Characteristics of the sites surveyed in southeastern Amazonia (see Figure 1).

Table 2.

Main results of primate surveys at the three study sites.

Figure 1.

The Xingu-Tocantins interfluvium in southeastern Amazonia showing the sites surveyed in the present study (see Table 1), and collecting and sighting localities of Callicebus moloch, according to: (a) Hershkovitz (1990), (b) Ferrari and Lopes (1990), and (c) Mascarenhas and Puorto (1988).

Discussion

The results of the surveys indicate that titis are naturally absent (or extremely rare) from extensive tracts of continuous terra firma forest within the Xingu-Tocantins interfluvium. Exact limits are unclear, but if the sites surveyed here are typical, the species may be absent from a large portion of the forest between the two rivers. Obviously, any estimate of population size based on the assumption of a homogeneous distribution within this area would require substantial revision. There are few clues to the determinants of the observed pattern, although it seems likely that ecological factors are involved, as in the case of the silvery marmoset, Mico argentatus, a second species with a heterogeneous distribution in this interfluvium (Ferrari and Lopes 1990, 1996). In this case, however, M. argentatus is present at site 1, but is absent from Tucuruí, where Callicebus is apparently abundant. So, even if similar ecological factors are involved, they clearly have different effects on the distribution of the two species.

One pattern apparent from the distribution of sites (Fig. 1) is a possible association with riparian habitats, given that most of the localities are distributed along the Tocantins. This may be a sampling artifact, but it could also reflect specific habitat preferences. As the Tocantins has a long history of human colonization, it may even be that C. moloch prefers anthropogenic habitats over pristine terra firma forest. This may include the ECFPn, where human impact is negligible. Interestingly, Wallace et al. (1998, 2000) identified a possibly similar situation in eastern Bolivia, where Callicebus donacophilus was absent from large preserves of primary terra firma forest close to the Brazilian border, but abundant in the anthropogenic landscape further south.

While Callicebus moloch is not under any immediate threat of extinction at the present time, the results of this study highlight a number of potential problems for the conservation of both this and other Amazonian primates. To begin with, the species is apparently absent from two of the sites (1 and 3) with the best potential for the long-term conservation of the region's primates (Ferrari et al. 1999).

One other key site is the Carajás complex (Fig. 1), which includes a fully-protected area (the 103,000-ha Tapirapé Biological Reserve), national forests and indigenous lands (Companhia Vale do Rio Doce 2007). Callicebus moloch is known to occur in this area, although little is known of its exact distribution and abundance, and Toledo et al. (1999) report that a population of the exotic Callicebus brunneus may have been established in the area, derived from animals released from captivity. Otherwise, the interfluvium is characterized by widespread deforestation, promoted by a number of “mega-projects”, including the Caraj´s Mining Project, existing (Tucuruí) and planned (Belo Monte) hydroelectric dams, and the Trans-Amazon highway (BR-230), which bisects the region.

On a broader scale, these findings underscore a problem that may become increasingly important as the Amazon basin is colonized. Many species—not only of primates, but also of many other groups of organisms—are known from a very small sample of localities within an apparently vast geographic range. In most cases, it is assumed that the species occupies all the available habitat between localities or river barriers (Ayres and Clutton-Brock 1992), but there is increasing evidence of major lacunae in the distribution of many species (Ferrari 2004), a prime example being the red howler (Alouatta seniculus) in southwestern Amazonia (Peres 1997b; Iwanaga and Ferrari 2002). In the case of one other eastern Amazonian endemic, the Ka'apor capuchin (Cebus kaapori), an apparently very patchy distribution, combined with extremely low population densities almost certainly determined the delay in the discovery of the species until the end of the twentieth century (Queiroz 1992). These same characteristics have also contributed decisively to the current status of C. kaapori as one of the most endangered of Amazonian primates (Ferrari and Queiroz 1994; IUCN 2006).

Clearly, more reliable data are needed for many, if not most species of Amazonian primates, especially those with relatively large geographic ranges. In the meantime, it may be necessary to revise conservation parameters for some species, including their status, where estimates are based on potentially problematic data. As shown here, such caution may be especially important for the planning of protected areas.