Introduction

Predation, and the threat of predation, are considered to be important factors in structuring various aspects of primate life, including social organization (Cheney and Wrangham 1987; Cowlishaw 1994), foraging strategies (Boinski et al. 2000; Barnett et al. 2012a), sleeping sites (Isbell 1994; Quinn and Cresswell 2004; Barnett et al. 2012b), and time budgets (Heymann 1990; Stanford 2002; Ferrari 2009; Mourthé and Barnett 2014). But even though predation events are short, infrequent, and likely to be influenced by the presence of observers, they are witnessed by biologists (Boinski et al. 2000; McKinney 2009) and have provided a range of information on anti-predation strategies (e.g. Galetti 1996; Bezerra et al. 2009; Calquhoun 2007; Fichtel 2007; Gursky-Doyen and Nekaris 2007; Ferrari 2009; Ribeiro-Júnior et al. 2016; Suscke et al. 2016; Teixeira et al. 2016). Reports even of failed predation attempts are useful for broadening understanding of this aspect of primate biology (Ferrari 2009; Suscke et al. 2016), as is knowledge of how the targeted primates react to such attempted predation events.

In this paper we document eight attempted attacks and one successful attack by raptors on red-nosed cuxiú, Chiropotes albinasus (I. Geoffroy and Deville 1848), a medium-sized (3–4 kg) arboreal Neotropical pitheciine primate, and provide information on the anti-predation strategies of the cuxius. The red-nosed cuxiú is endangered and occurs south of the Amazon River and east of the Rio Madeira, in southeastern Amazonia (Pinto 2008). They inhabit a variety of forest types, including never-flooded terra firme rainforest, seasonally flooded forests on river margins, and dry scrub forest (Wallace et al. 1996; Ferrari et al. 1999; Pinto 2008).

The red-nosed cuxiú is poorly known compared to other members of its genus (Veiga and Ferrari 2013). Only three ecological studies of 6 months or more have occurred (Ayres 1981; Pinto 2008; Soares da Silva 2013), though the species has been recorded by a number of short-term studies and surveys (Ayres and Milton 1981; Branch 1983; George et al. 1988; Ferrari 1995; Wallace et al. 1996; Ferrari et al. 1999, 2003; Vaz 2001; Rocha and Silva 2013). From these studies information has been gathered on red-nosed cuxiú diet, distribution, habitat preference, and group size.

To our knowledge, this is the first time predation and anti-predation strategies have been recorded for Chiropotes albinasus, though there are reports of predation on C. chiropotes (Lenz and dos Reis 2011: harpy eagle attacking a juvenile female), and C. utahicki (Ferrari et al. 2004, Boa constrictor attacking an adult female; de Souza et al. 2005, harpy eagle attacking an adult male).

Methods

The data reported here were collected during four independent studies, and at four different localities (Fig. 1). In all cases, predation was recorded as part of standard field protocols.

Fig. 1
figure 1

Locations of the four studies, in relation to the geographic range of the red-nosed cuxiú (IUCN 2015), the Amazon basin (Lehner et al. 2006), and the major rivers and their tributaries in South America (Lehner et al. 2006). Study 4 consisted of a survey that covered 400 km, and it covered the area denoted by the locations of 3 and 4 on the map

Full size image

Study sites

Site 1: FLONA do Tapajós, Pará State, Brazilian Amazonia

A 17-month study (August 2004–January 2006) of the foraging and social ecology of red-nosed cuxiú (Pinto 2008), occurred at the Floresta Nacional do Tapajós (FLONA do Tapajós), a 527,000 ha sustainable-use area within a much larger block of lowland terra firme forest, situated on the east bank of the Rio Tapajós, a south-bank tributary of the Rio Amazonas. The red-nosed cuxiú study focused on the area near the Sucupira Study Base (03°21′19″S, 54°56′57″W). The study used a trail system in a 1400 ha grid surveyed for 10 days every study month. Using the instantaneous scan sampling method (Altmann 1974) at a 10-min interval, the researchers recorded the behaviour and location of all individual cuxiús in sight. The researchers logged 556 h of direct field observation of the cuxiús.

Site 2: Cristalino Private Reserve, Mato Grosso state, Brazilian Amazonia

This 6-month study (March–September 2012, except August: 154 h of direct field observation) examined the effects of seasonality on diet and group size on red-nosed cuxiús (Soares da Silva 2013). The study area was a 6,593 ha fragment of lowland terra firme forest, sustained by commercial ecotourism, and located between Cristalino State Park and the Cristalino River (9°35′45″S, 55°56′14″W). Full-day follows of the red-nosed cuxiús occurred for 10 days per month for 6 months. Because of the presence of ecotourism groups in the reserve, the cuxiús were habituated to people. Using the instantaneous scan sampling method (Altmann 1974) at a 5-min interval, the researcher recorded the behaviour and location of every individual cuxiú in sight.

Site 3: Terra firme forests, Middle Rio Tapajós, Pará State, Brazilian Amazonia

This 21-month research project (April 2012–December 2013: total of 328 field surveys hours) was part of a broader biological inventory related to civil construction projects (de Oliveira et al. 2016). Using a number of field camps, researchers conducted in lowland largely pristine terra firme forest along eleven 5 km trails through the terra firme forests of the Middle Rio Tapajós (04°52′45″S, 56°26′45″W). Trails ran perpendicular to the river. Researchers surveyed the trails, at a slow regular pace, five times each during every survey period, so that trails were visited a total of 265 times during the whole study period, which accounted for more than 1325 km walked.

Site 4: Riverine forests of the Itaituba-Machado stretch of the middle Rio Tapajós, Pará State, Brazilian Amazonia

This 21-month research project (April 2012–December 2013: total of 328 field surveys hours) was part of a broader environmental impact analysis (de Oliveira et al. 2016). Researchers conducted surveys for primates in the seasonally flooded riparian forest (igapó: sensu Prance 1979) on river margins between the town of Itaituba (04°16′33″S, 55°59′02″W) to the rapids above the community of Machado (04°56′28″S, 56°45′22″W) and along both sides of the river Jamanxim to the first set of impassable rapids—a total river-travel distance of some 400 km. Igapó on the Tapajós is flooded for up to 6 months of the year. It forms a narrow band ca. 20 m wide that is botanically distinct in composition from adjacent terra firme (Ferreira and Prance 1998). Researchers surveyed the area during both the flooded and unflooded seasons, noting the size and composition of each group of primates.

Ethical note

All four studies were non-invasive studies, and had received the necessary permission and permits from the Brazilian government. All research adhered to the legal requirements of Brazil.

Results

In a total of 1255 h of fieldwork (710 h of direct observation of red-nosed cuxiús and 545 h of primate surveys), nine predation-linked events were observed. Of these, eight failed and one was successful. Four events involved harpy eagles (Harpia harpyja), two involved Leucopternis sp. hawks, and one involved a black hawk-eagle (Spizaëtus tyrannus). In two other instances the large raptor involved could not be identified. Attempts were directed at both adult and non-adult cuxiús and involved both adult and juvenile raptors. Attempts were made in both terra firme forest and in igapó.

Site 1: Flona Tapajós

Three predation-attempts were observed. All occurred in lowland terra firme rainforest, and none were successful.

  1. (1).

    On 19 Dec 2004 at 16:30 a medium-sized hawk landed close to a group of 20 red-nosed cuxiús foraging in the upper forest canopy. The animals dropped to mid-canopy and then ran and leapt quickly while making a coarse grunting alarm call. The response lasted 13 min, after which the cuxiús moved a further 500 m to a sleeping tree. The raptor involved was a species of Leucopternis. Both Leucopternis kuhli and L. melanops have been recorded in the region (Oren and Parker 1997; Barlow et al. 2002), but as bill colour and head-pattern were not seen clearly, Leucopternis sp. is used from here-on for such raptors.

  2. (2).

    On 16 Nov 2004 at 16:33, an adult harpy eagle made a predation attempt on a subgroup of 10 red-nosed cuxiús. The raptor stooped to catch a foraging adult, but missed when the animal ran and hid in dense foliage. After the incident, the cuxiús appeared vigilant and very nervous and immediately left the area, moving quickly through the mid-canopy for approximately 2 min until they reached two large trees. During this period the group was widely spread out (>150 m). On arriving at the large trees, the group positioned themselves in the lower part of the upper canopy so that the vertical trunks and large semi-vertical main branches of the two trees were between them and the raptor. They then moved, and seven individuals positioned themselves in the canopy of a large tree, the exterior of which had a dense covering of vines. The remaining three individuals positioned themselves some 15 m away in another tree with a very dense canopy. At this point, the eagle, which had been actively following and watching the group for approximately 12 min, left the area. Throughout this time, the group members were emitting intense whistles, combined with a very high-pitched whinny-like cacacacá–um–um. After the eagle left, the subgroup members reunited, with the inter-individual distances (1–2 m) being much less than those when they foraged (2–10 m: Pinto 2008). The sub-group moved quickly and in silence for 10 min until they arrived at a sleeping tree where they spent the night.

  3. (3).

    On 18 Nov 2005 at 16:40, an adult Leucopternis sp. attacked a group containing at least 10 adult red-nosed cuxiús by flying directly at one adult. As the bird did so, another adult cuxiú came running at high speed in the direction of the raptor. The raptor then left the area. It was not possible to determine the sex of the bird or primates involved. Over the next 10 min the cuxiús became very vigilant and emitted frequent alarm vocalizations (described above), as well as a great number of agonistic low-pitched whistles.

Site 2: Cristalino Private Reserve

A single predation attempt was recorded on an 18-member group (14 adult and four infant animals). At 09:30 on 1 May 2015, a young male harpy eagle stooped at a sub-group of seven red-nosed cuxiús. The sub-group contained four females with infants and three adult males. The attack occurred shortly after the sub-group had split from another seven-strong sub-group (all adult females without infants), from which they were separated by several tens of meters.

At the time of the attack the cuxiús were moving in upper canopy, returning to a point where they had been foraging previously. As the eagle approached, one group member made a very loud, scream-like call, and immediately all members leapt to the branches of the stratum below, where they remained, quiet, immobile, and concealed. Seconds later, the eagle landed close to where the primates had concealed themselves. It remained there for 16 min, during which time the cuxiús remained very quiet and concealed in dense vegetation. The eagle then moved approximately 150 m away and continued to observe the group, which remained silent and immobile until the eagle left the area 33 min later. The primates did not vocalize until 78 min after the eagleʼs departure.

Site 3: Terra Firme, Rio Tapajós

One observation of a predation attempt involving red-nosed cuxiú occurred between 08:30 and 09:00 on 12 Jul 2012. A large unidentified eagle repeatedly attacked a mixed group of red-nosed cuxiús and tufted capuchins (Sapajus apella). This mixed group of primates was in the dense canopy when the attack began. Although several strikes were made, none were successful. After the attack, individuals of both primate species descended to the understory and vocalized continuously for approximately 10 min. The mixed group then actively followed the researchers, at 2–3 m above the forest trail, for several minutes before moving up into the mid-canopy and moving off, still vocalizing.

Site 4: Riverside forests, Tapajós

There were four observations of predation-associated events involving red-nosed cuxiú in this habitat.

  1. (1).

    On 11 Oct 2012 at 14:48 an adult female harpy eagle was seen perched in the upper branches of a dead tree in igapó forest some 15 m behind a group of red-nosed cuxiús (six adults, three sub-adults, one dam-dependent juvenile) that was feeding in a grove of babaçu palms (Attalea speciosa). As the eagle approached the primate group, shrill whistles and two or three whinny-like alarm calls were heard, and much of the group dropped into dense vegetation below. A small group (including at least one adult male) remained in the palm canopy and the eagle dove at them twice before returning to its original perch. During this period, no alarm calls were heard, but twittering vocalizations began and continued for some 10 min.

  2. (2).

    On 23 Sept 2013 at 10:14, a large unidentified eagle flew close to a group of three to five adult red-nosed cuxiú entering the upper canopy of an emergent terra firme tree approximately 1 km from the researchers. The cuxiús plummeted vertically into the denser vegetation below. Given the distance, vocalizations, if emitted, could not be heard.

  3. (3).

    On 25 May 2013 at 17:17 in igapó forest, a sub-adult harpy eagle was seen eating a young red-nosed cuxiú. The bird was considered a sub-adult because the crest was whitish tipped with dark-gray/black (not fully dark grey or black as in an adult), the chest immediately below the throat was pale-grey with dark mottling (not solid dark-grey as in an adult), and the wings and back were also pale (not dark-grey as in an adult). It was likely that the bird was in its third year (harpies mature in their fourth year: Howell and Webb 1995). The primate was considered a young individual because it appeared to be approximately 2/3 the size of the bird: mean body length of an adult red-nosed cuxiú is approximately 42 cm (Hershkovitz 1985), and that of adult harpy eagle is 86.5–107 cm (Howell and Webb 1995). No other red-nosed cuxiú were recorded in the immediate region, but a series of their alarm vocalizations had been heard at a point 300 m upstream at 17:02. The bird appeared to be tearing at the chest region, the area of the prey which is often opened by puncture from the talons at the initial strike (Ferrari and Port-Carvalho 2003);

  4. (4).

    On 10 Oct 2013 at 08:13 in igapó forest, an adult black hawk-eagle flew at a group of six adult red-nosed cuxiús as they moved through the upper canopy of forest at the terra firme/igapó forest transition. The strike missed, but as the animals dropped to lower vegetation an infant became dislodged from its mother’s back. The infant scrambled onto a branch, concealed itself within a large bromeliad colony, and let out a piercing high-pitched vocalization. Three adult cuxiús (including at least two adult males) returned swiftly along branches and charged the eagle which had settled on an open branch some 4 m from the infant. The charge was accompanied by aggressive bark-like vocalizations, as well as the whinny-like calls and whistles previously heard in an alarm context. It was not clear which animals were making which calls. The eagle left and the infant joined the adults. The infant was then briefly groomed by them before they all left in the direction of the rest of the group. At this point the group was lost from sight, but twittering calls spread over a wide area continued to be heard for another 5 min before all contact was lost. The event lasted a total of 8 min, 30 s.

Discussion

There was no obvious association between the number of predation events observed and the duration of the study. However, predation was far more commonly observed in igapó (four events in 217 h) than in terra firme (five events in 1028 h) forests. While it is likely that greater ease with which primates can be seen in igapó may make it easier to record predation events occurring there, igapó on the Tapajós occurs as a very narrow band (Ferreira and Prance 1998) and for a few months of the year, primates are present at high densities in this habitat. The observed predation may represent highly opportunistic activity by H. harpyja and other raptors, given that the red-nosed cuxiú is generally a very infrequent item in harpy eagle diets in terra firme forests, where they form less than 1% of the recorded prey captures (Silva 2007; Aguiar-Silva et al. 2014).

Six main anti-predation behaviours were observed during the four studies: mobbing, alarm calling, fleeing, freezing, increasing of inter-individual distance within a group, and hiding. A combination of at least two of these behaviours was performed in each of the nine encounters reported. These behaviours have previously been observed in several primate species as aerial anti-predator strategies (Stanford 2002; Scheumann et al. 2007; Barnett et al. 2015). While Peetz (2001) reported that cuxiús react collectively towards predators, Ferrari et al. (2004) noted that, because all cuxiú species practice fission–fusion sociality, this response may vary with size of the group at the time of the attack. Our data agree with Ferrari et al. (2004); of the nine events recorded, the two involving active defence against a raptor occurred when groups were relatively large (N > 10 adults). While approaching a predator is clearly risky (Cowlishaw 1994; Rose et al. 2003; Tórrez et al. 2012), group defence and rescues are not unknown (Chapman 1986; Passamani 1995; Perry et al. 2003; Caro 2005), and even punitive attacks have been recorded (Boinski 1988). Though exceptions exist where both males and females actively engage in group defence (Matsuda and Izawa 2008; Mourthé 2011), male primates (sub-adult and adult) are commonly most active in mobbing, group defence, and rescues (Boinski 1988; de Luna et al. 2010; Shahuano-Tello et al. 2002; Tórrez et al. 2012). This involvement by males from the group was the case with at least one of the rescues that we observed. Similar to the findings of anti-predatory behaviour in the equatorial saki (P. aequatorialis) by de Luna et al. (2010), and as predicted by theory (Isbell 1994; Arlet and Isbell 2009), such defence was observed in the cuxiús when members of the most vulnerable cohort, non-weaned infants, were present in the group. It is notable that, in the incident observed at Cristalino Private Reserve, the attacked sub-group consisted of three adult males and four females with infants, and, while no active defence occurred, all the group males accompanied the infant-carrying females. The remaining seven adult females, which were separated from the rest of the group by some 30 m, continued alone.

Such post-attack spreading out is a strong anti-predator strategy as it widens the area from which the cuxiús can detect potential predators (while still being able to communicate vocally), facilitates individual concealment, and may confuse the predator, encouraging it to give up the attack (Ferrari et al. 2004). Post-attack bunching occurred when the initial group was small, and may function in group cohesion.

Following an attack, the most common post-predation response by primates is to move away in silence (e.g. Ateles belzebuth, Matsuda and Izawa 2008, Mourthé 2011; Chiropotes utahicki, de Souza et al. 2005; Cebus capucinus, Tórrez et al. 2012; Papio cyanocephalus, Matsumoto-Oda 2015; Sapajus xanthosternos, Suscke et al. 2016: though see Barnett et al. 2011; de Luna et al. 2010 for exceptions). Except for one event, all predation attempts were accompanied by cuxiú alarm calls, followed by silence and vigilance. The high-pitched whinny-like cacacacá–um–um vocalizations heard during attacks were recorded only under these circumstances. Such sounds are very similar to the vocalizations reported by Barnett et al. (2011) for Cacajao ouakary following a harpy eagle attack and more generally as an alarm call for the species (Bezerra et al. 2010). The red-nosed cuxiú is normally a very vocal species, with a repertoire of short, high-pitched contact calls that are uttered near-continually when the group is moving (Pinto 2008).

Moving into denser vegetation followed all of the unsuccessful predation attempts. Such activity either during or after an attack is a common primate reaction, has been recorded in Paleotropical and Neotropical primates, and is to provide increased protection and a base for elevated vigilance (Cäsar et al. 2012; Matsumoto-Oda 2015). Post-attack movement in silence to a sleeping tree was unusual for the cuxiús, as they normally vocalized until the last couple of minutes before entering a sleeping tree (Pinto 2008).

Unless prevented by habitat structure (Barnett et al. 2011), primates often drop into low vegetation or even travel on the ground following an attack by an aerial predator (Lledo-Ferrer et al. 2009; de Souza et al. 2005; de Luna et al. 2010; Barnett et al. 2015). Such responses may also occur when predation on another species occurs in close proximity (E. Setz, unpublished data, saw Pithecia pithecia drop into lower vegetation when an unidentified hawk tried to capture a silky anteater, Cyclopes didactylus, on a branch nearby). Dropping action was observed in seven of the nine events reported here. Under such conditions, animals may also spread out while moving (Gilbert 2000; de Souza et al. 2005; Miranda et al. 2006; Lledo-Ferrer et al. 2009), which was observed in two of the nine incidences reported here. When occurring after a predation attempt, when all group members are alert, such behaviour is believed to reduce the likelihood of new encounters with the predator (Frechette et al. 2014). In this context, it is notable that most attacks occurred early in the morning or late in the evening when the cuxiús are more active and can be located more easily by the predators.

Young primates may be selected as prey because they are less experienced and, being smaller, easier to capture (Gilbert 2000; Vasquez and Heymann 2001; Stone 2007; Urbani et al. 2012; Riley et al. 2015). In the incidences reported here, of the three individuals identified as definitive targets of an attack, two were adults, and one was a juvenile. In other instances, the target was not clear, but no juvenile animals were obvious.

Because predation is unpredictable and rarely observed, reporting is often opportunistic and fragmentary in nature. Yet it is important to gain information on predation as it (or the risk of it occurring) is widely considered to be one of the major factors structuring primate societies, including their group size and vigilance (Boinski et al. 2003), which in turn impact time budgets and govern how and where primates forage (Mourthé and Barnett 2014). Given the central importance of predation to primate behaviour and ecology, and the great need that still exists for sufficient data to appreciate the nuances of primate anti-predator behaviour, it is suggested that other researchers combine their rare, and often descriptive, observations of predation attempts. Studies focusing on quantitative descriptions of primate alarm calls (and playback experiments using those calls), could aid our understanding of the primate anti-predation system and help clarify the influence of predation and predation risk on the evolution of primate behaviour and communication.