Observing Animal and Human interactions around a Marburg Virus-infected Bat colony.

Zoonotic infections, diseases which spread from an Animal host into a Human population, are a serious threat to public health, in some cases being the source of pandemics which kill millions of people around the globe, such as the Influenza and Covid Viruses and the medieval Plague outbreaks. Filoviruses, such as the Bundibugyo, Ebola, Marburg, and Sudan Viruses, can cause outbreaks of hemorrhagic fever which kill hundreds or even thousands of people in tropical Africa, after jumping from a natural host, usually a Fruit Bat. The Bats themselves seem to suffer little harm from infection with these Viruses, but Humans and other Animals often succumb very rapidly. 

Since the nature of the wild host of these Viruses became apparent, public health campaigns in many countries have tried to minimise the extent to which people interact with Fruit Bats, but the often remote nature of the communities most at risk, combined with the ability of the Virus to jump from Bats into a variety of other hosts before infecting Humans, has limited the effectiveness of these, underlining the need to better understand the nature of interactions between Fruit Bats, Humans, and other Animals.

In a letter published in the journal Current Biology on 20 April 2026, Bosco Atukwatse, Orin Cornille, Johnson Muhereza, Winfred Nsabimana, and Yahaya Ssemakula of the Kyambura Lion Project of the Volcanoes Safaris Partnership Trust, Eric Enyel of the Uganda Wildlife Authority, Charlie Gould of the School of Biological Science at the University of Edinburgh, Arjun Gopalaswamy of Carnassials Global, and Alexander Braczkowski, also of the Kyambura Lion Project of the Volcanoes Safaris Partnership Trust, present the results of an study in which camera traps were deployed at Python Cave in the Queen Elizabeth National Park, the dwelling place of a colony of  Egyptian Rousette Bats, Rousettus aegyptiacus, known to act as a reserve for Marburg Virus.

The camera traps were deployed as part of a carnivore-monitoring program in Queen Elizabeth National Park between 16 February and 23 June 2025, with a total of 8832 hours of filming producing evidence of at least 14 species of Animal visiting the size, including Carnivores, Primates, Raptors, and Reptiles. The most common visitors to the cave were Nile monitors, Varanus niloticus, (72 visits), Large Spotted Genets, Genetta tigrina, (69 visits), and Palm-nut Vultures, Gypohierax angolensis, (35 visits). The cave was visited twice by Olive Baboons, Papio anubis, and ten times by Blue Monkeys, Cercopithecus mitis, which were seen actively preying on Bats. African Fish Eagles, Icthyophaga vocifer, made 33 visits to the cave, Black Sparrowhawks, Astur melanoleucus, made 19 visits, Crowned Eagles, Stephanoaetus coronatus, made 17 visits, and Verreaux’s Eagle-owls, Bubo lacteus, made six visits. Seventeen Leopard visits were directly recorded, but parts of a cycle in which the same Leopard repeatedly entered the cave, captured Bats, and exited with them were recorded, leading the team to conclude at least 43 such hunts took place during the study time. A total of 63 incidents of hunting or scavenging Bats by different species were recorded, with a further 258 incidents of Animals entering the caves for less clear reasons.

 Multi-species predation and scavenging at a Marburg Virus Bat reservoir. The distribution of detections per species across 368 trap nights at Python Cave recorded using  6 remote, solar-powered camera traps. Blue indicates predation and scavenging events, whereas  orange represents detections where cave exploration, entry, exit, or resting was observed. Atukwatse et al. (2026).

More alarmingly, over the course of the study Atukwatse et al. also observed 214 individual Humans approaching the cave. These came in 22 different groups, including school groups, researchers, and tourists, many of whom came within meters of the cave entrance (a direct contradiction of park regulations), bypassing a designated observation platform 30 m from the entrance. Only one person, a tourist, was observed to wear a mask when approaching the cave.

Atukwatse et al. do not suggest that these observations represent evidence of Virus transmission, but rather a direct record of ecological interactions at a potential spillover site. This included direct predation of potentially-infected Bats during repeated visits to the cave, and provide direct evidence of the predator-guild targeting Bats in this setting. 

A collage of hunting and foraging by a variety of species. Insets show confirmed prey contact and feeding events. (i) An African Leopard, Panthera pardus, with its Bat prey emerging from the cave interior. (ii) A Crowned Eagle, Stephanoaetus coronatus, with its Bat prey. (iii) A Blue Monkey, Cercopithecus mitis, holding a Bat in its left hand. (iv) A Melanistic Genet, Genetta victoriae, with Bat prey. (v) A Nile Monitor, Varanus niloticus, approaching and then consuming a fallen Bat. (vi) An interspecific interaction (likely a fight) between a Crowned Eagle and a Nile Monitor over two Bats captured by the Eagle. (vii) An African Civet, Civettictis civetta, scavenging on Bat remains. (viii) A Palm-nut Vulture, Gypohierax angolensis, scavenging a Bat carcass. (ix) A group of Olive Baboons, Papio anubis, at the cave mouth, possibly foraging on Bat guano. Atukwatse et al. (2026).

Unlike other famous tropical Bat roosts within caves, such as Kitaka Mine in Uganda, Kitum Cave in Kenya, Goroumbwa Mine in Democratic Republic of Congo, or Macaregua Cave in Colombia, Python Cave lacks a vertical space separating the Bats from ground-based predators. Parts of the cave roof have collapsed, and the cave is partly filled with large volumes of guano, providing predators a way of directly reaching the Bats where they roost. This it turn provides an easy opportunity for any Virus present in the Bat community to spread to other wildlife within the cave. Atukwatse et al. observed instances of bats falling from the overcrowded roof of the cave, then having to crawl over the floor, presenting another tempting target for both terrestrial and airborne predators.

First ever large scale predation by at least 14 species on Egyptian Fruit Bats (a known Marburg Filovirus reservoir) in Python Cave Uganda. The trail cameras were placed at the cave as part of the Volcanoes Safaris Partnership Trust Kyambura Lion Project's long term Leopard and Hyena monitoring work in Queen Elizabeth National Park. Alex Braczkowski/YouTube.

The presence of piles of Bat bones and frequent and repeated visits to the caves by a variety of predators indicates that the site has become a prime feeding site for both predators and scavenges, and that to some extent the richness of the resource has led to a relaxation of normal inter-species hostilities, with, for example, Fish Eagles feeding alongside Vultures, Nile Monitors feeding alongside Pythons, and even a Genet and Python seen together. This implies a loss of territoriality and aggression only seen at the most abundant food sites, although on one occasion  Atukwatse et al. did observe a Crowned Eagle and a Nile Monitor squabbling over a Bat.

These interactions are all the more remarkable in that they are occurring at a site with a frequent Human presence, something which many Animals avoid where possible. The Uganda Wildlife Authority has placed an observation platform about 30 m from the entrance to Python Cave, with the specific intention of limiting Human exposure to Marburg Virus-infected Bats, but despite this Atukwatse et al.'s camera traps recorded Tourists, students from a nearby wildlife training institute, and even school trips approaching the cave mouth without any form of protection. It is particularly concerning that this was happening during the Bat's birthing season, when they are known to shed the Virus at a higher rate. Atukwatse et al. recognise that Bat-viewing is a valuable contribution to the nation's ecotourism income, but nevertheless recommend that the Uganda Wildlife Authority imposes stricter regulation on the site, with mandates for protective gear, enforced distancing, and locally trained guides to serve as sentinels for biosurveillance and education.

It has generally been assumed that the spillover events which lead to outbreaks of Marburg Virus Disease, and other zoonotic infections, occur in locations effectively beyond the observation of science. Atukwatse et al.'s results refute this, providing an example of a site where Marburg-infected Bats are interacting with a variety of other Animals, which is also integrated into the local tourist industry. Some of the species observed, such as Blue Monkeys, have previously been associated with Marburg Virus, and are commonly hunted and consumed by Humans as bushmeat. The consumption of Bats by these Monkeys presents a new route by which Marburg Virus could make its way from the wild Bat reserve into Human populations. Atukwatse et al. recommend that serological surveys are carried out of both the predators seen frequenting Python Cave, and of park rangers who regularly enter the site, in order to assess potential exposure to Marburg or other Filoviruses. This, combined with enhanced surveillance of Python Cave and other similar sites, could enable the development of a new dataset to complement the genomic tools already being developed.

See also...

Sibling rivalry in wild Chacma Baboons.

Jealousy occurs in Humans when we perceive our relationship with a person important to us is threatened by the actions of someone else. This is a complex emotion, which often manifests in attempts to disrupt social interactions that we find threatening. Whether jealousy is a uniquely Human emotion is less clear, as it can be hard to judge the genuine emotional state of non-Human Animals, at least in part due to the dangers of anthropomorphising behaviours that resemble Human emotional traits.

Non-Human Primates would appear to be a good starting point for those wishing to study jealousy outside our species, but very few studies appear to have been done in this area, and all of those on sexual jealousy in captive Primates held in unnatural conditions.

In Humans, one area which has been extensively studied is sibling rivalry among children. Human mothers are unusual in that they often care for multiple offspring of different ages at the same time, leading to forms of competition for parental attention and resources which do not occur in most Animals. Intense rivalries can develop between child siblings when they perceive that they are being treated differently, sometimes leading to conflicts which extend into adulthood. Such rivalries are most likely to develop when children are the same sex and close in age.

Such rivalries between siblings of different ages have not been studied in non-Human Animals. Instead, research has concentrated on competition between siblings produced in large clutches or broods, where they are part of a group reliant on parental provisioning until reaching independence. There have also been some studies concerning mother-child competition for resources in monotocous species (species that have one young at a time), plus a few studies looking at species in which older siblings cooperate in the rearing of young. 

Monotocous Animals with long child-rearing periods, including many Primates, form an intermediate group between Humans and polytocous species (species that have produce many offspring in a single clutch or litter). In such species different aged siblings interact with one-another and their parent(s) on a daily basis, and competition between siblings has the potential to have a significant impact upon fitness. Siblings will compete with one another in ways which they do not compete with non-siblings, requiring maternal care and attention rather than just foraging competitively. 

In a paper published in the journal Proceedings of the Royal Society B on 11 February 2026, Axelle Delaunay of the Institute of Evolutionary Science of Montpellier, and the Tsaobis Baboon Project in the Tsaobis Nature Park in Namibia, Vittoria Roatti, also of the Tsaobis Baboon Project, and of the Department of Anthropology at University College London, Rose Ellis and Punaete Kandjii, again of the Tsaobis Baboon Project, Alecia Carter, again of the Tsaobis Baboon Project and the Department of Anthropology at University College London, and of the Gobabeb Research Institute, Guy Cowlishaw of the Tsaobis Baboon Project, the Gobabeb Research Institute, and the Institute of Zoology, Marie Charpentier, also of the Institute of Evolutionary Science of Montpellier, and Élise Huchard, once again of the Institute of Evolutionary Science of Montpellier, the Tsaobis Baboon Project, and the Gobabeb Research Institute, present the results of a study into the role of jealousy in driving sibling competition in wild Chacma Baboons, Papio ursinus, a monotocous, social Primate.

Chacma Baboons live in matrilineal societies in which males disperse to new groups around puberty, while females remain within their birth groups, inheriting their mother's dominance rank. On average, females produce one young every two years, with growing Baboons are weened after about two years, but have a long developmental period and have long-lasting strong bonds to their mother, expressed through grooming relationships in which preference of grooming partner is an important signifier of relationship status.

In order to do this, they looked at instances in which young Baboons interrupted their mother when she was grooming a sibling. They did this to test three hypotheses, that young Baboons interrupted on such occasions because they were jealous, because they wished for attention from their mother themselves, or because they wanted to interact with their sibling. Theoretically, a jealous Baboon would interrupt even if it was not likely to gain any reward, a Baboon wanting attention from its mother would only interrupt if this was likely to result in it getting such attention, and a Baboon wishing to interact with a sibling would only do so if it was likely to gain such interaction. 

Theoretically, a Baboon which was jealous or wanted to play with a sibling would be more likely to approach its mother when she was grooming that sibling, while a Baboon simply wanting its mothers attention would be most likely to approach her when she was unoccupied. Furthermore, the jealousy hypothesis suggests that a Baboon would be more likely to interrupt its mother if she was grooming a sibling which tended to monopolise her attention, or otherwise appeared to be a favourite. This would fit with the patterns observed in Human children. 

It was also predicted that, in Baboons, jealousy would be more prevalent among same-sex siblings, and in particular between sisters, since younger sisters can come to outrank older sisters with maternal support. Since males leave the troop at puberty and do not rely on their mothers for social status, they were predicted to be less prone to jealousy. 

Furthermore, it was predicted that Baboons seeking maternal care would be more likely to interrupt when their mother was grooming a younger sibling, as these tend to be easier to displace, and that females may be more inclined to interrupt when their mother was grooming a male sibling, as mother-daughter social bonds are stronger than mother-son bonds.

If Baboons interrupt grooming because they want to play with a sibling, then males interrupting when other males were being groomed would be predicted to be the more frequent occurrence, as young males play with other young males more frequently than females play with females, or young Baboons play with members of the opposite sex. This would also lead to more interruptions when the Baboon being groomed was close in age to the Baboon interrupting. 

Dalauney et al. studied Chacma Baboons from two well habituated troops (L and J) in the Tsaobis Nature Park, which lies on the edge of the Namib Desert in Namibia, between August and December 2021. The Baboons were followed from dawn to dusk each day by trained observers who were able to identify all Baboons in the troops, including infants. Every family group which included at least a mother and two offspring was included in the study. This comprised eight families in each troop, with between two and five offspring, with a total of 28 female and 21 male young Baboons, ranging in age from six days to 8.9 years. This included adult females, but not adult and subadult males still living in the maternal group, due to the limited amount of interaction these males have with their mothers. 

Female Baboons were observed for five minute intervals when they were either grooming one of their offspring, or resting, and all interruptions were recorded, as well as whether this was be a sibling or non-sibling of the Baboon being groomed. Interruptions could be aggressive, such as attacking, chasing, biting, pushing, slaping, threatening, displacing, or supplanting; affiliative behaviours such as body contact, come-here faces, grunting, jumping on one groomer, lipsmacking, playing, presenting, or touching; tantrum behaviours such as gecks, complaint grunts, or other screams; maternal care solicitations, such as soliciting access to the nipple, soliciting grooming, suckling, or starting a triadic grooming session with the groomers; or simply approaching within a metre of a grooming pair. Under this analysis, a very broad range of behaviours were treated as 'interruptions', the object being to understand the motives of young Baboons in approaching their mothers. As such, focusing only on agonistic behaviours or effective disruption would only provide a partial picture of sibling interference, and might not pick up on some relationships, for example if younger siblings were afraid to behave aggressively towards older siblings being groomed by their mothers, but still found other ways to gain her attention. 

Having come up with a definition of 'interruptions', Dalauney et al. then looked at the outcome of these interruptions, grouping them into three basic categories; instances where the grooming continues, instances where the grooming stops, and instances in which the interrupter replaces one of the Baboons in the grooming relationship. In the final case, they recorded which Baboons were in the new grooming relationship (i.e. mother and interrupting youngster, or interrupting youngster and previously involved youngster). Finally, Dalauney et al. recorded every youngster within 10 m of their mother at the start of a five minute session. If the mother broke off from her original activity (grooming or resting) to engage in some other activity during the five minutes of the observation, that observation was abandoned. Each family group was monitored for no more than five minutes each hour. 

A juvenile male just approached and initiated body contact with his younger sister being groomed by their mother. This interference did not interrupt the ongoing grooming interaction, nor allow him to groom with his mother or his sibling. Axelle Delaunay in Dalauney et al. (2026).

Dalauney et al. found that young Baboons were significantly more likely to interrupt their mother when she was grooming a sibling than when she was resting, which predicted by the theories that the young Baboon was jealous or wanted the attention of its sibling, but not by the theory that the primary motivator was gaining the attention of the mother. This was unaffected by the presence of other siblings within 10 m (which was predicted to make interruptions less likely if a young Baboon was looking for a sibling to play with), nor was it affected by the sex of the youngsters, nor the social status of those involved. Baboons did become less likely to interrupt their mother grooming a sibling as they got older.

Younger Baboons were more likely to interrupt their mother when she was grooming a sibling, but this was apparently unaffected by the age-difference between them and the sibling being groomed. However, older siblings were twice as likely to interrupt their mother when she was grooming a younger sibling, and all Baboons were more likely to interrupt their mother when she was grooming a sibling of the same sex, with males more likely to interrupt when a brother was being groomed than females were to interrupt when a sister was being groomed. If one of the mother's offspring was perceived as a favourite, and received more grooming attention than its siblings, then those siblings were more likely to interrupt when it was being groomed. The age of this favourite did not appear to matter. This again supports the theory that interrupting Baboons were driven primarily by jealousy. 

Of 501 instances of Baboons interrupting their mother while she was grooming a sibling, 95 instances (19%) resulted in grooming being broken off (the predicted desired outcome of the jealousy model), while in 44 cases (9%) resulted in the interrupter entering into a grooming session with the mother (the predicted desired outcome of the maternal attention-seeking model), and only 12 (2%) resulted in the interrupter entering into a grooming session with the sibling (the predicted desired outcome of the sibling attention-seeking model). Entering into a grooming relationship with the mother or a sibling did not necessarily disrupt the original grooming relationship, as Baboons can form grooming triads. Data was not initially collected on interrupters entering into play relationships with the sibling being groomed (another possible outcome of the sibling attention-seeking model), but only 11 instances of this were observed, limiting support for this hypothesis. Thus, while most attempts at interrupting a mother grooming a sibling failed to disrupt that activity, they were significantly more likely to cause that grooming session to break off than to allow the interrupting Baboon to enter into a grooming relationship with one of the originally involved Baboons, supporting the hypothesis that the main reason for such interruptions was jealousy.

Interference are rarely successful, but more often disrupt the ongoing grooming than they grant access to maternal or sibling grooming. Proportion of successful interference as defined under the jealousy hypothesis (i.e. the interference effectively disrupts the ongoing mother–sibling grooming, in pink), the care-seeking hypothesis (i.e. the interferer gains access to maternal grooming, in orange), and the sibling-seeking hypothesis (i.e. the interferer gains access to sibling grooming, in yellow). Note that interferer do not need to disrupt the ongoing grooming to gain access to maternal or sibling grooming, as interference leading to triadic grooming interactions was also considered successful. The blue bar represents all the interferences that were not successful under any hypothesis, i.e. when the grooming kept going and the interferer did not get access to either of the groomers. Dalauney et al. (2026).

Dalauney et al.'s study lends to support to the idea that young Baboons interrupt their mothers when they are grooming siblings out of jealousy. They were more likely to interrupt their mothers when they were grooming siblings than when she was apparently available. They were also more likely to interrupt when the sibling being groomed was younger than them, of the same sex, or perceived as their mothers favourite. This interference appeared to be largely aimed at the mothers rather than their siblings. In Baboon societies, mothers are a more useful social connection than siblings, as they provide support to female offspring throughout their lives, and to male offspring for as long as they remain within their maternal group, whereas brothers will leave the group at some point, and sisters will shift the focus of their support to their own offspring when they start to have them.

Younger siblings were preferentially targeted over older siblings. This may be because younger Baboons tend to monopolise their mothers attention, or because they are less likely to respond aggressively to such interruptions. There is a lack of comparative studies of this behaviour in Humans, where only the jealousy of older siblings towards younger siblings has been explored, and then only in the context of age difference between siblings in modern Western societies.

What has been recorded in both modern Western societies and non-Human Primates is that behaviour predicted to be caused by jealousy is more commonly directed at younger siblings, and siblings of the same sex (which implies that in both cases offspring of different sexes are competing for slightly different maternal resources, and that siblings of the same sex are therefore a greater threat). In Baboons and Humans, young males engage in more rough-and-tumble play than young females, which might lead to a desire to play with siblings becoming a reason to disrupt interactions between those siblings and their mothers. Dalauney et al. believe that by looking at a wide range of interrupting actions and their outcomes, that their study has screened for this possibility

Dalauney et al. believe that the tendency of Baboons to interrupt when their mothers are grooming a sibling perceived as a favourite is important. This matches the finding in Humans that children who feel disfavoured by their parents report higher levels of conflict with both parents and siblings, often with lifelong consequences. In Humans, siblings are less likely to be jealous if they perceive that siblings receiving different treatment are doing so because they have different needs. Non-Human Primates are known to be able to track relationships between their close family and social partners, as well as those between third parties. Dalauney et al.'s findings suggest that Chacma Baboons are able to judge the strength of their maternal bond compared to that of their siblings, although the cognitive ability needed to do this and the perception of fairness in Baboons will need to be the subject of further studies. Negative responses to uneven food rewards have previously been recorded in several Primate species, as well as Domestic Dogs and Corvids. Whether Primates can assess maternal care in the same way is still unclear, but the ability to do so would clearly have implications for family dynamics.

One of the most important questions that arises from Dalauney et al.'s study is whether there is any benefit to a young Baboon in simply disrupting its mother grooming a sibling, and whether there is any more general advantage to jealousy. Interrupting activity was shown to be twice as likely to result in the mother abandoning grooming of a sibling as it was to result in the interrupter entering into a grooming session with the mother, and entering into such a grooming bout does not actually require the mother to abandon the other youngster, as Baboons can form grooming triads. Notably, the most common outcome was that nothing changed, i.e. the mother continued to groom the sibling she had been grooming when interrupted. This suggests that any benefits associated with jealousy may be complex and time-delayed, particularly if young Baboons are able to manage their emotional state sufficiently to refrain from an activity unlikely to pay off. Any adaptive benefits of jealous behaviour are likely to change over a lifetime, which should also alter their expression. There is less advantage to a Baboon monopolising its mothers time and attention as it becomes older and more independent, able to find its own resources and make its own social connections. Under these circumstances, sibling connections might become more important than maternal connections. In strictly evolutionary terms (i.e. the passing on of genes), it may be advantageous to an older juvenile Baboon to stifle jealous instincts in order to let a younger sibling have more time with its mother.

Ultimately, whether we are Human or Baboon, our emotions are a private matter, confined within our heads, which others can only judge by our actions. It is therefore possible that, when judging the emotions of another species, that we anthropomorphise their motivations, and provide emotional explanations for actions which have been determined by entirely different biological pathways. However, there is a growing body of evidence coming from a wide range of taxa, including Primates, Ungulates, and Corvids, that individuals will try to interfere with a close partners interactions with others, apparently in order to protect a valuable social relationship. Most studies have not attempted to evaluate the emotional underpinning of such actions, but these responses clearly correspond to what we would call jealousy in Humans. Dalauney et al. suggest that the role of emotions in managing social relationships in non-Human Animals has been underestimated, despite the fact that the primary function of emotions is to provide a way for individuals to act upon external stimuli. Jealousy is a social emotion, something which helps us to navigate complex social environments, and comparisons of Human and non-Human responses to similar social stimuli can potentially unlock ways to understand the emotional lives of Animals. 

See also...

Lethal conflict during the fission of a Chimpanzee group in Kibale National Park, Uganda.

Warfare, and other forms of collective violence, are one of the more distinctive, if less pleasant, forms of Human behaviour. In these acts, Humans are able to sort themselves into groups defined by traits other than kinship, such as ethnicity, religion, language, or other cultural traits, in order to engage in conflict against other groups of Humans. It has been proposed that this is a result of our ability to use cultural traits to define cohesive social groups, giving us a distinct sense of who is 'us' and who is 'them', enabling closer cooperation with members of our group, while at the same time maintaining a hostility to outsiders. 

However, this cannot explain how conflict can erupt within formerly cohesive groups, leading to rebellions and civil wars. This has been explained by suggesting that shifting interpersonal ties and rivalries can sometimes reach a point where internal hostilities overcome the cohesion of the group, regardless of the cultural connections which have been built up. This hypothesis has some evidential support, with observations suggesting that Humans can rapidly come to regard former members of an in-group as members of an out-group for the most arbitrary of reasons. 

Non-Human Animals also engage in territorial aggression, and sometimes lethal conflict with cospecifics, despite not having the religious, cultural, or political ideologies seen in Humans. This has been observed in a variety of Animals, including Banded Mongooses, Lions, Wolves, and Primates. Non-Human Animals living in social groups also have episodes of fission, in which one group splits permanently into two or more, something which is often explained in terms of feeding competition and social pressures, although most Animal groups do not engage in lethal combat during such fission episodes.

Chimpanzees, Pan troglodytes, are among our closest relatives, and in some ways can be seen as a bridge over the gap between Human and non-Human behaviour. Male Chimpanzees stay within the group they were born in, and engage in cooperative defence of the group's territory, as well as raids on the territories of neighbouring groups. Lethal violence between males of different groups is quite common during both of these activities. This can be explained in terms of hostility towards outsiders, while remaining loyal to members of the kin group. Chimpanzee groups are known to occasionally split, a process which is thought to involve lethal violence between males familiar with one-another.

In the 1970s, a group of Chimpanzees living in the Gombe National Park in Tanzania was thought to split in two, with the males of one of the new groups subsequently killing one of the adult females and all six adult males in the other group over a period of about four years. However, this group was not under constant observation, so that much of this activity is inferred rather than having been directly witnessed, and no subsequent observations of similar splits in Chimpanzee groups were recorded in the following decades. Furthermore, a study published in 2014 which looked at the genetic structure of Chimpanzee groups suggested that such ruptures were extremely uncommon, with groups splitting on average once every 500 years.

In a paper published in the journal Science on 9 April 2026, a group of scientists led by Aaron Sandel of the Department of Anthropology at the University of Texas at Austin, present a detailed and extensively documented study of a fission event which split a group of Chimpanzees living in the Kibale National Park in Uganda, and the lethal violence associated with this split.

Sandel et al. analysed 24 years of data on the social interactions of the group, ten years of GPS data, and 30 years of demographic data on the Ngogo Chimpanzee Group, which demonstrated a three step process, in which a formerly cohesive group polarised into two clusters with little social interaction between these clusters, these clusters then actively avoided one-another for two years, before engaging in a series of lethal aggressive actions. Sandel et al. take this as evidence that Chimpanzee groups can fracture and engage in collective violence against members of the same community without any of the cultural markers seen in Humans.

The Ngogo Chimpanzees have been the subject of a continuous research project since 1995. For the first two decades of the study, they remained a single group, although they did go through a regular fission-fusion dynamic in which the larger group split into parties which changed throughout the day, though individuals frequently moved between parties and all remained within the overall territory; this is something commonly seen in Chimpanzee groups. Females typically migrated from (or to) the group at adolescence, while males born into the group remained there for life. The males formed a strict dominance hierarchy, associated in mixed-sex parties, hunted together, and cooperated in territorial patrols.

Each adult male in the group was followed for 2-3 months each year between 1998 and 2024 to see which other individuals they associated with, stayed close to, and engaged in grooming with. Despite being a single group, the Ngogo Chimpanzees typically split into two-to-four clusters over the course of a year, with two persistent and long-lived clusters, the Western and Central clusters. Membership of these clusters was fluid, with 29% of Chimpanzees switching cluster each year, and extensive ties maintained between Chimpanzees in different clusters. 

As well as clusters, Sandel et al. identified a number of 'cliques' of males that consistently stayed together, even when switching clusters. One of these cliques comprised a group of three males that would go on to form the core of the post-fission Western Group. These three males remained together consistently, even when forming clusters with males that would go on to be in the post-fission Central Group. A cluster comprising exclusively males that would go on to form the post-fission Western Group first appeared in 2014.

In 2015, Ngogo Chimpanzees at the edge of their territory hear calls from neighbouring Chimpanzees; adult males embrace each other in an act of reassurance before moving toward the calls and engaging in an intergroup encounter. 'C' and 'W' designations refer to males who would later become members of the Central and Western groups, respectively. Sandal et al. (2026).

Despite the different social clusters, all of the Ngogo Chimpanzees, including all of the males, had overlapping space use patterns, and all males shared the same set of reproductive partners. All Chimpanzees born within the Ngogo Group between 2004 and 2014 that it was possible to genetically sample had both parents from within the group, though 44% had parents from two different clusters. 

The first sign of a split between the clusters was observed on 24 June 2015, when members of the Western and Central clusters were seen to approach one-another near the centre of the territory. Unexpectedly, rather than the two groups merging as usually happened on such occasions, the Chimpanzees of the Western Cluster ran away, with the Chimpanzees of the Central Cluster chasing them. The two clusters then avoided one-another for six weeks, something which had never been observed before.

The events of 2015 precipitated the greatest change of social change seen throughout the decades-long study of the group. What had been a single large group of Chimpanzees split into a number of smaller units, stabilising as two new groups by 2018.

The first patrol by one group against another happened in 2016, when males of the Western Group, accompanied by two of the Central Group males, staged a patrol against the Central Group. All subsequent patrols by the Western males contained only members of that group. In 2017, the Central Group males staged their first patrol against the Western Group, with aggressive interactions between the two groups escalating rapidly from that point. During one encounter in 2017 the males of the Western Cluster attacked the alpha male of the Central Cluster (who had been part of the Western Cluster before 2014), severely injuring him. Both groups subsequently increased the number of patrols against the other group.

By 2017 the two groups were using largely distinct territories, with the overlap between the two groups being similar in size to that seen between unrelated groups. The centre of the shared territory had become a border. This was accompanied by reproductive isolation between the two emerging groups, with the last infant with parents from different groups being conceived in March 2015. All subsequent births had both parents from the same new group.

By 2018, the original group appeared to have split into two new groups, entirely separated from one another. At this time the Western Group included 10 males and 22 females over the age of 12, and the Central Group comprised 30 males and 39 females over the age of 12. Until 2018, a few of the females and infants from the Central Group would occasionally join the Western Group when foraging from Fig trees. After 2018 all such activity ceased.

Following this complete split, the Western Group initiated a series of lethal attacks against members of the Central Group. On all occasions members of the Central Group were attacked by multiple members of the Western Group, during a patrol by members of the Western Group into the territory of the Central Group. Six lethal attacks on males belonging to the Central Group were observed between 2018 and 2024, with a seventh attributed with a high degree of confidence. From 2021 these attacks were also carried out on infants from the Central group, with fourteen infanticides observed, and another three inferred. 

Sandal et al. note that this is a conservative estimate. Between 2021 and 2024 another 14 adult and adolescent males from the Central Group disappeared, and are thought likely to have been killed. None of these individuals showed signs of any illness when they were last observed.

In 2019, the Western Chimpanzees, including W1, W2, and W3, attack and kill the adult male Central Chimpanzee C1, who they had associated with before the fission, marking a notable shift in relationships. Sandal et al. (2026).

By using decades of data gathered on the Ngogo Chimpanzees in Kibale National Park, Sandel et al. have been able to demonstrate the occurrence of a fission event within a wild group of Chimpanzees, something predicted to happen only once every 500 years. This was followed by a series of lethal aggressive interactions, with targeted violence continuing years after the split, something not observed in any other non-Human Primate. The rate at which killings occurred was far higher than that seen in small Human societies. This demonstrates that Chimpanzees can develop new group boundaries and defend them violently, despite not having any concept of ethnicity, religion, or political ideology.

Sandal et al. speculate that a number of factors might have contributed towards the division of the Ngogo Chimpanzee Group. Firstly, the group was unusually large, with almost 200 individuals, and almost 30 males, much larger than other known Chimpanzee groups, which may have strained the males ability to maintain good relations with all other members of the group. Secondly, feeding competition has been shown to play a role in group fission in other Primate species. The area occupied by the Ngogo Group had abundant food, but the large size of the group may have caused strain at some times of the year. Thirdly, the two groups became reproductively isolated before finally separating into two groups, something which may have increased male-male aggression as they had to compete for a smaller number of mates.

Other factors which may have played a role are also observed by Sandal et al.. Firstly, six adult members of the group, five males and a female, died in 2014. The cause of these deaths is unknown, but two showed signs of illness before their deaths, making it possible that the group was hit by a disease. The loss of more than 10% of the male Chimpanzees in the group may have weakened the groups network of social relationships, leading to the group beginning to break up in 2015. 

Next, the dominant male in the group changed in 2015, immediately before the fission of the group. Such changeovers are known to raise tensions among male Chimpanzees, leading to increases in behaviours such as aggression and avoidance. The former dominant male belonged to the Central Group, whereas the new dominant male came from the Western Group, but moved to the Central Group when he ascended to the top of the hierarchy, which Sandal et al. suspect may have increased tensions between the two groups.

Finally, in January 2017, the group was hit by a respiratory epidemic which killed 25 Chimpanzees, including four adult males and ten adult females. Two of the males that died were from the Western Group, including one of the last males in that group to be maintaining relations with the Central Group. Thus, even though this event happened after the groups had started to split, it may have contributed to the final breakdown in relations between the two. 

Chimpanzees are known to have a strong sense of who is in their group, and who is not. Female Chimpanzees leave the group they were born into as adolescents and look for a new group to join, but males remain in the group they were born into their entire lives. Thus any unfamiliar male Chimpanzee is treated as a stranger, with no recorded instances of wild male Chimpanzees forming relationships or cooperating with Chimpanzees in other groups. Under these circumstances, intense hostility towards outside males appears to be an adaptive trait in male Chimpanzees, leading to potential territorial extensions for the group, which in turn leads to more food and other resources, increased female fertility, and a greater survival rate among juveniles. 

This cannot, however, explain the lethal aggression sometimes displayed to members, or former members, of the same group. During the fission of the Ngogo Group, males which had lived, fed, groomed, and patrolled together for years became targets of lethal attacks on the basis of their new group membership. This leads Sandal et al. to conclude that Chimpanzees have a sense of who belongs to their group which is based upon more than simple familiarity, and which can be updated to reflect changes in circumstances.

These results challenge previous assumptions about intergroup conflict in Chimpanzees in a number of ways. All observed attacks were initiated by the numerically smaller Western Group, which contradicts the assumption that larger groups have an advantage in such conflicts. The emergence of greater social cohesion in the smaller group appears to have given them an advantage in conflict which more than made up for the greater numbers of their opponents. These closer bonds appear to have developed before the onset of aggressive activity, suggesting that an external threat is not needed to forge these bonds. Conversely, the original Ngogo Group underwent a territorial expansion in 2009, which appears to have reduced the threat that it faced from rival groups, something which may have contributed to the long-term decline in social cohesion within the group. If this is the case, than an external threat may not be needed to forge close bonds, but it may be needed to maintain them.

Sandal et al.'s findings also have implications for Human behaviour. If Chimpanzee groups can split in this way without any of the cultural markers associated with Human conflict, then these may be less important to Humans than we generally assume, masking the actual roots of aggression, which may have more to do with interpersonal relationships. Humans tend to attribute the conflict to ethnic, religious, or political divisions, but this may be misleading, covering the actual causes of conflict. If this is the case, then  it may be in the small, daily acts of reconciliation and reunion between individuals that we find opportunities for peace.

See also...

The distribution and conservation status of Wild Rooibos Plants, Aspalathus linearis.

Rooibos Tea is a traditional herbal drink from South Africa, not marketed internationally. It is made from the leaves of the Rooibos Plant, Aspalathus  linearis, a type of Cape Gorse which grows wild in dry areas of the northwestern part of the Fynbos Biome, part of the wider Cape Floristic Region. Although traditionally harvested from wild plants, the modern Rooibos industry is built around large scale cultivation of the plant, with only a tiny amount of the total crop harvested from the wild.

Thus the Rooibos Plant is considered to be an economically significant plant in South Africa. Although the commercial harvest is tied to agricultural production, this the wild population preserves a  wider genetic heritage, which may become important should the commercial crop be threatened by disease or climate change. However, while quite a lot of research has been published into Rooibos, most of this relates to commercial cultivation, leaving us with little understanding of the ecology, distribution, or conservation status of the wild plants.

In a paper published in the South African Journal of Science on 26 November 2025, Tineke Kraaij of the Natural Resource Science and Management Cluster at Nelson Mandela University, Vernon Visser of the Centre for Statistics in Ecology, Environment and Conservation, and National Institute for Theoretical and Computational Sciences, at the University of Cape Town, and Gerhard Pretorius of NaturaLibra Environmental Services, present the results of a study which looked at the current distribution, potential distribution, population size, and threats to the population size of Wild Rooibos Plants in South Africa, with a view to creating a baseline against which future monitoring of the species could be measured.

Rooiboss is grows in areas with cold wet winters and hot dry summers, with annual rainfall levels of between 300 and 600 mm, and at elevations of between 450 and 900 m. It is found as far north as Vanrhynsdorp and as far south as Betty's Bay and the Cape Peninsula, and anecdotally grows as far east as Riviersonderend and Bredasdorp, 

Wild Rooibos is an extremely variable species (potentially being a species cluster - a group of closely related species difficult to tell apart - rather than a true species), and can be difficult to differentiate from closely related African Gorse species such as Golden Tea, Aspalathus pendula, or Lebeck's Roiibos, Aspalathus lebeckioides. Wild Rooibos also has several different ecotypes, strains of the plant with different morphologies and environmental preferences, further complicating efforts to understand the distribution of the species.

The ecotypes of Wild Rooibos vary in habit, fire-survival strategy, vegetative and reproductive morphology, biochemistry, and genetics. They can be loosely grouped into four main forms, 'Erect' or 'Upright', 'Prostrate', 'Bush' or 'Shrub', and 'Tree' (although 'Tree' is sometimes regarded as a variant on the 'Erect' or 'Upright' form). A 'Salignus' form is also sometimes recognised, this having a Willow-like growth with multiple lanky stems reaching 2-3 m tall and a sparse canopy. The Prostrate and Shrub forms are wider than they are tall, and tend to have many branches close to the ground, and can resprout after a fire. The Erect, Tree, and Saliginus forms are tall, and only regrow from seeds. In addition to these, there is the commercially grown 'Nortier' form, a cultivar bred from the Erect form, which is tall like its parent, but also has the dense branching of the Bush form, giving it a large harvestable biomass. 

Exemplary images of the different ecotypes or growth forms of Wild Rooibos, Aspalathus linearis, namely the (a) Bush or Shrub form, (b) Prostrate form, (c) Erect (including Upright or Tree) form, and (d) Salignus form. Kraaij et al. (2025).

These different ecotypes have different distributions. The Bush, Prostrate and Tree types occur towards the northern part of the species’ range, and the Prostrate, Erect and Salignus types toward the south. The Erect and Salignus types favour higher elevations, above 400-600 m. The Bush ecotype is predominant in areas with low rainfall (less than 200 mm per year), while the Salignus ecotype favours wetter areas (more than 500 mm per year), and lower elevations. Different ecotypes often coexist in the same region, but seldom close to one-another. 

Kraaij et al. built up a map of the distribution of Wild Rooibos and its various ecotypes, using data from the Global Biodiversity Information Facility and CapeNature databases, previously published research, and fieldwork carried out by Gerhard Pretorius. As much as possible, sites with historic records of the plants were visited to establish their continuing presence (or absence). Additional observations were made along roads connecting historic observation sites.

This data was then used to build a model of the potential distribution of the species, and the Bush, Prostrate, Erect, and Saligus ecotypes. This was achieved by mapping the presence of the plants against a range of variables, including precipitation in the  warmest quarter; precipitation in the coldest quarter; maximum temperature of the warmest month; minimum temperature of the coldest month; topographical slope; soil electrical conductivity; soil potassium, sodium, phosphorus, carbon, and nitrogen levels, soil pH; and the proportion of sand in the soil. 

Of the 235 sites where the species was historically recorded, it was found to now be absent from 30, and present at 47. In addition, 99 new locations where the species was present were identified. The species was not found at an outlier north of Nieuwoudtville and one east of Vanrhynsdorp, and some locations around Citrusdal, not was it found south of Malmesbury, at Franschoek, on the Cape Peninsula, or in the Gordon’s Bay and Betty’s Bay areas. It was confirmed at De Doorns, the easternmost location where it was previously recorded. Kraaij et al. note that the absence of the species at a site where it was historically recorded does not imply local extinction; it is possible that the previous records were miss-identifications. Notably, the habitat at some of these locations appeared completely unsuitable for Rooibos, with very steep slopes, clay soils, or wetlands, and a plant assemblage different from that associated with Rooibos plants. Some records of the plant in southern areas listed on iNaturalist could be verified, suggesting that it is not absent from the claimed southern extent of its range, but the species mainly occurs in the area between Nieuwoudtville in the north, Groot Winterhoek in the south, Wupperthal and the Tra-Tra mountains in the east, and just southeast of Graafwater in the west.

Results of field verification of historical location records of Wild Rooibos, showing where Rooibos was present, absent or not verified, as well as new location records. Kraaij et al. (2025).

The Wild Rooibos population was predicted to extend from Nieuwoudtville in the north to just southeast of Ceres, with the bulk of its distribution around Clanwilliam and Citrusdal, which was largely in line with predictions based upon previous reports. In total the species occupied an area of about 6000 km², with individual ecotypes occupying areas of between 1500 and 2500 km². The Salignus ectotype was the most abundant, predicted to occupy an area of about 254 600 km², followed by the Prostrate ecotype, 161 200 km², the Bush ecotype, 121 100 km², and the Erect ecotype, 119 900 km². Because each ecotype and the species as a whole were modelled separately, the species total does equal the total for each ecotype minus the overlap between different ecotypes. 

Overlap between ecotypes was only substantial in the core area of the distribution range. The Bush and Prostrate ecotypes had roughly similar distributions, extending east and south of the core area. The Erect and Salignus ecotypes had distributions extending west and north of the core area, with the Salignus ecotype predicted to extend furthest west, beyond Piketberg. The Salignus ecotype was also predicted to be found in the south, around Ceres. 

The distribution of the different ecotypes appeared to be driven by different factors. The distribution of the Bush and Prostrate ecotypes appeared to be influenced by the coldest cool period temperature, with both ecotypes able to cope with sub-zero temperatures. Warm season precipitation was clearly important to these ecotypes as well, with both favouring low summer precipitation. The distribution of the Erect and Salignus ecotypes was most strongly influenced by the concentration of potassium in the soil, with both ecotypes favouring very low concentrations. The Erect ecotype also appeared to need low soil conductivity. The distribution of both Erect and Prostrate Ecotypes was also influenced by the concentration of carbon in the soil with both ecotypes favouring very low levels. Kraaij et al. note that high winter temperatures and wet summers favour the growth of pathogenic Oomycetes such as Pythium spp. and Phytophthora cinnamomi, which are known pests of farmed Rooibos.

Kraaij et al. also produce estimated population sizes for the species and each ecotype. They again note that the total for the species does not match the total for each ecotype added together, as these were calculated separately. Three estimates for each population were made, based upon the minimum, mean, and maximum density of the plants. For the species as a whole, the minimum population was estimated at 14.7 million plants, while the mean was 1.78 billion plants, and the maximum was 17.0 billion plants. The most abundant ecotype, Salignus, was estimated to have a minimum population of 144 million plants, a mean population of 390 million plants, and a maximum population of 630 million plants. For the Prostrate ecotype the figures were, minimum 4.35 million plants, mean 863 million plants, and maximum 4.66 million plants. For the Bush ecotype the figures were minimum 13.3 million plants, mean 325 million plants, and maximum 1.34 billion plants. Finally, for the Erect ecotype, the figures were minimum 3.00 million, mean 189 million, and maximum 833 million plants.

Kraaij et al. recommend, however, that these figures be taken with a good deal of caution, as characterising the species distribution and abundance in the landscape was challenging in many cases, with significant implications for the population estimates. The density of Prostrate plants was particularly hard to estimate, particularly when it was interspersed with other plants. The species has a highly irregular distribution across its range, with plants often forming dense clumps in the midst of apparently similar and equally habitable, but unoccupied, landscapes. Both the size and density of individual populations varied a great deal from site to site. Furthermore, most populations were detected from a moving vehicle, requiring a minimum density for observation; it is possible that less dense populations were overlooked. Finally, Rooibos plants grow in a landscape prone to periodic burning. This does not appear to damage the population as a whole, but the plants do take time to recover after a fire, so their presence is likely to have been missed in areas which had recently burned.

Given these constraints, Kraaij et al. estimate that the populations of the species and each ecotype are probably towards the lower ends of their estimated numbers, but that more detailed surveys of each population would be needed to confirm this.

Much of the population of Wold Rooibos appears to be on protected land. Kraaij estimate that 27% of the total population is found on state or private protected areas, with 38% of the Prostrate ecotype, 33% of the Salignus ecotype, 28% of the Bush ecotype, and 20% of the Erect ecotype occurring on protected land.

About 40% of the Wild Rooiboss population was found to be growing in areas where Kraaij et al. considered that it faced no threats. Another 13% was found in areas where it was prone to illegal harvesting, although this was not thought to present a major threat to the species. About 28% of the population was growing in areas where land transformation presented a threat, although much of this land was probably unsuitable for conservation, and was likely to receive some protection from legislation preventing the clearing of natural vegetation. About 4% of the land where the species grew is potentially threatened by overgrazing, although this does not appear to present a direct threat to Wild Rooibos plants. About 15% of the sites were threatened by invasive plants, particularly Dodder, Cuscuta campestris, a parasitic climbing plant from North America, which can target a wide range of hosts and which is a problematic invasive species in several parts of the world. Cultivated Rooibos fields are known to be prone to invasion by Golden Wreath Wattle, Acacia saligna, and Cyclops Wattle, Acacia cyclops, but while these are a problem in Rooibos fields, they do not seem to be a problem in the natural setting. Potentially, climate change could alter the fire regime in the areas where Wild Rooibos grows. However, the species currently grows in areas where fires are fairly infrequent, with the interval between fires typically significantly longer than the regrowth of the plants, so any shift in fire regime would have to be significant to cause a problem.

Climate change could also potentially alter rainfall patterns and temperature in areas where Wild Rooiboss grows. Since these are variables known to affect the distribution of the plants, this could potentially have a major impact on the species, particularly if winters become warmer and wetter. Two of the populations visited during the study appeared to have suffered widespread mortality events following high summer rainfall events in the summer before being visited. Cultivated Rooibos is also known to be intolerant of waterlogged soils, requiring farmers to implement careful drainage schemes. Since a wetter climate is a predicted outcome of a warming climate in the Western Cape, this could be a major threat to the future of the species. 

Wild Rooiboss does not appear to be at risk from hybridisation. The various ecotypes do not appear to interbreed, even when in close proximity to one-another. Potentially each ecotype could be threatened by inbreeding depression if populations become isolated. As it stands, geneflow within the species is not really understood, though it is likely that geographical barriers such as wide rivers and mountain ranges present a barrier to pollination, and the main distributors of the seeds of the plants are Ants, which do not tend to move long distances. Despite the variety of ecotypes, Wild Rooiboss does not show a great deal of genetic diversity as a species, with even less diversity in the southern part of its range. This lack of diversity could potentially present a threat to the species, and should be considered carefully during any conservation efforts. 

There are more than 9000 fields of Cultivated Rooiboss in South Africa, with 4181 occurring within the natural range of the species as calculated by Kraaij et al. Of the 146 Wild Rooibos populations varified, 55 were found to be within 1 km of Cultivated Rooibos fields. Theoretically, hybridisation between Wild and Cultivated Rooibos is possible, but this was not observed, Cultivated Rooibos plants do not persist well after cultivation ceases, which, combined with the apparently limited gene flow between ecotypes, suggests that hybridisation with the Cultivated form does not present a major threat to Wild Rooiboss populations. 

Wild Rooibos has a limited and fragmented distribution, but within this range appears to be common. The total wild population of the species is estimated to be more than 15 million plants. About 27% of the species distribution lies within protected areas. The species does not appear to be threatened by illegal overharvesting, invasive plants, overgrazing, or fire regimes, nor by hybridisation with the Cultivated form. The potential for habitat transformation appears to present only a minor threat to the species. By far the greatest threat to Wild Rooibos appears to be the potential effects of global warming, potentially bringing warmer, wetter conditions unfavourable to the species. This could well cause the range inhabited by the species to contract, with little potential for migration to better climates without Human intervention. 

See also...