Cooperative rescue of a juvenile capuchin (Cebus imitator) from a Boa constrictor

The threat of predation by snakes is considered to have played a significant role in the evolution of primate sensory systems and behavior. However, we know relatively little about individual and group responses given the rarity of observed predation events. Here we report an observed (filmed) predation attempt by an adult Boa constrictor (~ 2 m) on a juvenile white-faced capuchin (Cebus imitator) in the Sector Santa Rosa of the Área de Conservación Guanacaste, Costa Rica. The snake caught the juvenile monkey on the ground during a terrestrial play session. When the victim screamed, the alpha male, alpha female, and another adult female ran to the scene, physically attacked the snake (with bites and hits), and pulled the victim to safety. Most group members participated in the vocal mobbing of the snake both during and after the attack. Based on the outcomes of this predation attempt and published reports of other B. constrictor attacks on primates, the coordinated efforts of ≥ 2 group members is needed for a successful rescue. This observation adds to our growing knowledge of cooperative group behavior and its importance in predator defense.

Predation is considered a major selective force leading to group living in many animal species 1 , including primates [2][3][4] , though documented predation attempts are rare. Isbell 5 suggested that the threat of predation specifically by snakes played a crucial role in shaping primate behavior and sensory systems. Given the wide geographic overlap between primates and snakes, it is not surprising that primates also developed an array of anti-predator behaviors that increase snake detection, such as vigilance behaviors, and deter predation, such as predator-specific alarm calls 6 and predator mobbing 7 . This latter tactic involves animals attacking a predator vocally (repeated alarm calls) and/or physically (e.g., hitting, biting, or throwing items at a predator) rather than running away 7 . In many species, mobbing behavior is cooperative (involving multiple group members), alerts other individuals to the location of predators, and can drive predators away and discourage future hunting in the area 8 . Though effective in many cases, predator mobbing, particularly with non-ambush predators, does have potential costs as it can result in the injury or death of individual participants 9 . Mobbing ambush predators, such as constricting snakes, is a relatively common behavior for primates as it is much less risky and the hunting success of ambush predators is greatly reduced once they are detected 7 . Here we report on an observed predation attempt by a Boa constrictor on a juvenile white-faced capuchin (Cebus imitator) and the subsequent mobbing and rescue by group members.
Despite the assumed threat snakes pose on primates, there are few published observations of predation events. Based on these limited reports, the consequences of snake predation events are usually fatal. Of course, the efficacy of defensive tactics will differ when individuals are attacked by venomous or constricting snakes. Attempting to rescue a group member is more dangerous if a snake is venomous, as it could easily envenomate and kill multiple individuals and there is little an individual and/or their group members can do to save an envenomated individual. Four attacks on wild primates by venomous snakes have been observed and published, with the victim dying in all instances [10][11][12] . However, in the case of attacks by constricting snakes, escapes can occur with the aid of group members, as the risk of engaging a constricting snake is much lower once the snake begins coiling around the victim. Indeed, of the 16 documented predation events on primates by constricting snakes (14 Boidae, 2 Pythonidae), seven reported interventions involving one or more group members physically interacting with the predator ("rescue attempts") ( Table 1). Five of these interventions resulted in the successful release and survival of the victim. Note that four additional predation attempts were documented, but we are excluding these from analyses as the observers did not witness the initial attack and response Here, we add to these limited observations by reporting an attempted predation, and successful rescue, of a juvenile white-faced capuchin (Cebus imitator) by a Boa constrictor in the Sector Santa Rosa (SSR) of the Área de Conservación Guanacaste, Costa Rica. Our team was filming a terrestrial play session prior to the predation attempt, thereby enabling us to capture the details of the group response to the attack on film. The capuchins at SSR encounter snakes at a rate of 2.85/100 h of observation and cooperative mobbing of snakes is common 13 . Usually, several group members will cooperate in the mobbing by directing loud alarm calls and/or threat vocalizations towards the predator while engaging in threatening facial and branch shaking displays [14][15][16][17] . However, they do not usually come into direct contact with snakes and maintain a safe distance while threatening the predator (Fig. 1). Given the potential costs of directly engaging with a deadly snake 7 , and the rarity with which this type of behavior has been observed and/or reported, additional observations are critical for cross-species comparisons on variation in individual and group responses to predation. Such data will collectively advance our understanding of snake predation pressures and their selective role in the evolution of primate sociality and cooperative behavior.

Methods
The SSR white-faced capuchins have been under intensive investigation since 1983, and members of multiple study groups are individually known and habituated 18 . The capuchins in SSR reside in groups ranging from 5 to 35 members and include multiple related females, multiple immigrant males, and their offspring 18 . SSR comprises approximately 100 km 2 of deciduous tropical dry forest located 30 km south of the Nicaraguan border. SSR experiences distinct wet and dry seasons, with nearly all of the annual rainfall (avg. 1,792 mm) occurring in the wet season between mid-May and mid-November 18 . The predation attempt reported here took place on July 17, 2019, during the yearly veranillo, the characteristic dry period that occurs for several weeks during each wet season in July and/or August 19 . Our observations were made on the Los Valles (LV) group, which has been the focus of intensive research from 1991-present. At the time of the observation, group size was 25, including 4 adult males (≥ 10 years), 8 adult females (≥ 6 years), 4 large juveniles (4-6 years), 4 small juveniles (1-2 years), and 5 infants (≤ 1 year). White-faced capuchins are largely arboreal, though they spend considerable time both foraging and playing on or near the ground 20 .
The video was captured on an iPhone XS Max at 1080p HD at 30 frames per second (fps) (see https ://drive .googl e.com/file/d/1sJeN ZkgZ7 iffzP Hz_4Gsr0 MrwtV PxbKl /view?usp=shari ng). Video recording (by SEL) began 0:28 s prior to the first scream by the victim, marking the beginning of the attack. The video was analyzed using Adobe Premiere Rush to slow down the footage to 20% speed during the moment of attack, release, and www.nature.com/scientificreports/ aftermath. Additionally, we analyzed the video in QuickTime using the frame by frame function. Figure 2 is a screen shot from the video.

Results
Intermittent video recording of the group had been occurring for two hours prior to capturing the boa attack on video on July 17, 2019. The entire duration of the B. constrictor (boa herein) attack on the victim was less than 20 s, measured from when the screams started (0:28:82 in the video) to when the victim was released (0: 47:19). Information on the key participants in the predation attempt and rescue are presented in Table 2. At 11:26 am, an unidentified large juvenile (~ 6 years) was caught by an approximately 2 m long boa while engaged in a terrestrial play session. The individual ID is based on the victim's body size in comparison to the known monkeys involved in the rescue. We were unable to definitively ID the victim due to the fast paced action of the attack, the obscured view in the video, and the victim's quick flight from the scene following its release from the snake. Here we provide a time-stamped account of the events. While we made every attempt to ID all of the individuals involved, even with reviewing the film frame by frame, we were unable to confirm all IDs due to the fast action of the event. We provide individual IDs only for those animals that we were able to distinguish with 100% certainty. 0:00:00-0:28:81 The majority of group juveniles and several adults were engaged in a play session that involved reciprocating chases, running along the ground, and jumping up and down between tree trunks, vines, and the ground. 0:28:82-Screams of the victim are heard and the camera is directed towards him/her. The victim's screams can be heard throughout the duration of the attack. 0:29:88-First snake alarm call is sounded; adult male PW, who was sitting on the ground 1.5 m from the boa (separated by vines and green vegetation), jumped from the ground up to a branch about one meter above the snake and began to alarm call while facing the snake.   . Alpha male HP (center back) biting the boa, adult female TH (center front) actively pulling the victim from the snake, and three group members alarm calling at the scene. Note: many more were in attendance but out of the image frame. www.nature.com/scientificreports/ Note, shown in the linked video: at 2:03:09-Adult female CH, three adult males (HP, HG, and BY) and an unidentified large juvenile actively threatening from a tree directly over the snake as they watch it move off and settle under some foliage and branches about 3 m away from the attack location. The snake remained motionless in this location, although it was still somewhat visible. The group remained in the area and continued to alarm call and direct threats and branch shaking towards the snake, well past the end of the video at 3:39 s. Approximately 20 min after filming ended, the group moved off to forage and our team followed the monkeys, leaving the snake behind.
Following the attack, several drops of blood were observed on the ground where the attack took place. Though we cannot know for certain that the blood was from the boa, there was blood visible on the snake following the bites by HP, who had fresh blood on his left hand and shoulder immediately following the event. HP showed no sign of injury after the rescue or in the days following the attack, thus we suspect that the blood on HP was from the snake. All other group members appeared to be unharmed, as none showed any injuries or issues with movement following the event.

Discussion
Cooperative predator defense has been observed in a variety of animals, including small birds (Hirundo rustica 21 , Malurus coronatus 22 , Manorina melanocephala 23 ), lobster (Panulirus argus 24 ), giant otters (Pteronura brasiliensis 25 ), and meerkats (Suricata suricatta 26 ). However, cooperative rescues from predatory attacks, like the case we describe here, are rarely observed in wild vertebrates and to our knowledge, have only been reported for primates (Table 1; see also Ref. 27 ), humpback whales (Megaptera novaeagliae 28 ), banded mongoose (Miunfos mungo 29 ), and possibly dolphins (Delphinidae 30 ). These rescue behaviors are considered a special form of cooperation as they involve one or more individuals putting themselves at risk to aid another, with no guarantee that the outcome will be successful, and no direct gain for the rescuer(s) 31,32 .
Our observations described here clearly highlight the efficacy of cooperative predator mobbing to rescue a group member. Similar to Perry et al. 14 , we observed that with combined effort in the trees and on the ground, the group surrounded the snake and attacked from multiple angles while simultaneously pulling the victim loose. When taken together, these two observations provide strong evidence that rescue behaviors are part of the normal behavioral repertoire for C. imitator. This cooperation renders it difficult for the snake to defend itself and continue to hold the victim. In examining the published reports on constricting snake attacks on primates (Table 1), the coordinated attack by multiple group members to rescue a groupmate appears to be the key to a successful rescue. In all five successful snake predation rescues, the victim was released only after multiple group members (≥ 2) physically attacked the snake. In the three unsuccessful rescue attempts, the physical attack by group members was minimal, with one case involving only a single bite to the snake from one individual 33 , another case in which a single female hit the snake 4-10 times with her hands 34 , and in the final case a single individual jumped on the snake intermittently for 30 seconds and was joined by a second individual for the last few jumps 35 . In the four remaining cases of constricting snake predations for which sufficient observations were made, the group response was limited to alarm calls and the victim did not survive [36][37][38][39] . Similarly in the constricting snake predation attempt on a sifaka described by Burney 40 , group members only sounded alarm calls and observers were certain the victim would have perished without their intervention. While the number of recorded constricting snake attacks on wild primates remains low, collectively they illustrate that group mobbing behavior accompanied by physical aggression and the combined efforts from multiple individuals can be effective in saving group members.
Why individuals cooperate has been a central question in behavioral ecology, particularly in cases where they risk their own lives to help others. Indeed in 2005, the evolution of cooperative behavior was identified as one of the "top-25 big questions facing science over the next quarter century" 41 . Kinship certainly promotes cooperative behaviors (e.g. 42 ), and in the boa rescue described by Perry et al. 14 , the victim's mother was the main participant in the physical attack of the snake along with the group's alpha male. Though we were unable to definitively identify the victim of the attack in our observation, we can limit the possibilities to just three large juvenile group members. Two of these three were closely related to the three adult females who most actively participated in the rescue.
While we cannot determine if kinship was a motivating factor for the adult females who rescued the victim, we know that none of the possible victims were related to the alpha male (he was not a member of the group during the period when any of the potential victims were sired). In the event described here, the alpha male was the most active participant in the defensive attack. It is doubtful that the victim would have escaped without his intervention, similar to the rescue reported by Perry et al. 14 . Within seconds of the victim's screams, the alpha male ran to the scene and physically attacked the snake while a low-ranking, subordinate male (PW), who was right next to the boa when the attack occurred, only alarm called. Though anecdotal, these observations support the broader trend of alpha males as principal protectors of their groups. Alpha male capuchins are the most vigilant group members and the most active during interactions with predators and extragroup conspecifics 43,44 . They also maintain testosterone levels that are significantly higher than subordinate males [45][46][47] . These extreme concentrations are not necessary for reproductive function, as subordinate males with lower concentrations are capable of siring offspring 48,49 . These high levels of testosterone likely enable alpha male capuchins to remain alert and quickly to respond to threats 55,56 , which in this case the alpha male did even prior to siring offspring in the group.
Observations of group responses to predation events, though rare, clearly support the hypothesis that predation has been a strong selective force driving sociality in primates (e.g. 50 ). While Isbell 5 argued that the coevolution of snakes and primates have strongly influenced the evolution of primate visual systems, observations of cooperative group rescues of victims from constricting snakes further supports the strong role these predators www.nature.com/scientificreports/ have had in shaping primate behavior and sociality. Given that such cooperative rescues have now been reported for platyrrhines, tarsiers, and strepsirrhines but not in cercopithecines or hominoids, the large bodied primates, indicates that the threat of constricting snakes may have been a particularly strong selective force in early primate evolution when primates were small bodied and, therefore, more susceptible to fall prey to constricting snakes.

Data availability
All data relevant to this publication have been presented or supplied in the pubished manuscript.