So much of our life is influenced by who is in our social networks: we rely on extended families, friends of friends of friends, co-workers and their connections to gain intelligence on everything from what books to read to how to vote to which jobs to pursue. But we are by no means alone in this reliance: social networks also affect the daily experiences and, indeed, survival of individuals in many animal species. That chimpanzees and other primates have complex social lives has been well known for decades. More recent studies have revealed that the actions of single birds, dolphins and other creatures make complete sense only in their social context. These findings could affect everything from conservation efforts to understanding our own social networks.
And the research on animals—which often uses techniques developed to study human group behavior—may provide feedback to inform further investigations by us and about us.
Why network analysis is needed
It took both time and new ways of thinking for students of animal behavior—ethologists—to realize just how important social networks can be in the animal kingdom.
On supporting science journalism
If you're enjoying this article, consider supporting our award-winning journalism by subscribing. By purchasing a subscription you are helping to ensure the future of impactful stories about the discoveries and ideas shaping our world today.
In the 1930s future Nobel laureate Konrad Lorenz published his now famous studies describing imprinting in geese—the instinctive emotional attachment of a newborn to the first caretakers it encounters during a critical period in development. Soon the idea that most creatures are basically robots, engaging in hardwired, programmed behavior (that is, under the control of genes) became dogma.
Quickly, however, researchers realized that external factors interacted with the underlying genetic programming. Nature (genes) plus nurture (environment) drove animal behavior. Although that statement may seem comprehensive, it is actually not terribly useful—nature plus nurture includes virtually every possible influence one can imagine.
Investigators thus began to examine how trial-and-error learning also shaped behavior. Along with the observations of field researchers, these studies forced the recognition that animals were much smarter than we gave them credit for: chimps and crows make and use tools; parrots solve problems using logic; elephants disable electric fences by dropping large rocks on them. In the course of studying such obvious intelligence, researchers also began to observe that some animals in groups learned behaviors by copying their group mates. And a particular group member might notice that it was being watched by others trying to glean information.
Of course, as physicists know, once you get beyond a two-body problem, things can get exceedingly complicated. Early attempts to study the ways that individuals in a social group interact thus tended to concentrate on interplay involving two or three individuals. Scores of studies focused on an animal copying the mate choice of another, on a group member spying on the fighting abilities of a potential competitor or on a scrounger stealing food from a more productive group member. But the more that ethologists studied such behavior, the more they realized that these interactions among a few individuals were just a hint of the intricate set of relationships among all the members of a group.
What was needed for a fuller understanding of the social life of animals was the recognition that many animals, just like we humans, are embedded within complex social networks—the relationships that connect each individual to every other group member.
How it is done
Modern application of this approach began in earnest about 15 years ago, when ethologists started to freely adopt methods long used by social scientists for the study of human social networks—first in workplaces or neighborhoods, later in virtual communities such as Facebook and Twitter.
Social networks in animals range from simple associations involving only a few individuals, such as a loose shoal of fish traveling together, to far more complicated configurations, such as might be found in a troop of baboons, where individuals are embedded in multiple overlapping affiliations (such as mating, dominance or grooming networks) that can influence group members both directly and indirectly. Networks can change frequently: members may come or go, and individuals can change their position and connections in response to disease, acquisition of knowledge and previous interactions.
In both simple and complex animal societies, interactions among network members have important implications for survival and reproduction. Accuracy of information about food, predators and mates, as well as the speed at which that information travels in a group, depends on a social network's structure. Who plays with, fights with or helps whom also depends on network structure. And diseases and parasites can be transmitted from one individual to another, without the two ever meeting, by passing the pathogen through intermediates.
As part of their overall assessments, researchers specifically identify several features of animal networks: the keystone individuals (those that have many connections and whose removal disrupts the social network); nodes (any individual that is included within the network); density of the network (a ratio of the number of actual ties to the number of all possible ties); degree (the number of ties between each individual and all others); reach (the number of friends of the friends of an individual); and centrality (the percentage of all connections among individuals that include a given individual). For example, most people in the U.S. have low centrality at the scale of country, but with almost everyone aware of the president and connected to him by their local officials, his centrality approaches 100 percent.
To get a feel for how social networks operate in nature—and how they may be the key driver of the ways that everyone in the group ultimately behaves—let us peek into the not so private lives of three nonhuman species that have been examined in this way.
Macaque police keep the network intact
Pig-tailed macaques (Macaca nemestrina) create multiple networks, such as ones composed of playmates or grooming partners. Networks differ in size, and a monkey might have different favorite partners in different networks. A particular macaque may also play a stronger role in one network than in another. But the various networks share a common feature: they operate under the watchful eye of a few authority figures that keep the peace. These macaque “police,” some of the group's highest-ranking males, spend time and energy breaking up fights between other individuals in their social networks.
Jessica Flack of the Sante Fe Institute and her colleagues (including renowned Emory University primatologist Frans de Waal) studied the role of the police in a troop of 84 macaques at the Yerkes National Primate Research Center at Emory in the early to mid-2000s. Geneticists often decipher the role of a single gene in a cell or an organism by disabling the gene and observing the consequences of its absence. Flack's team adapted this “knockout” approach to the macaques by removing three policing males. They then watched and waited.
The loss of a low-ranking group member did little to the social networks. But as might be expected, the absence of police led to increased aggression and decreased reconciliation after fights in the population. Less predictably, without the police present, the play and grooming networks also underwent complex restructuring.
With police gone, for instance, group members played with and groomed fewer partners; that is, the “degree” of their play and grooming networks decreased. And the “reach” of the remaining monkeys—the number of friends of the friends of an individual—went down in those networks. At the same time, the cohesion of the entire society weakened; the population underwent a kind of balkanization, dividing into smaller, more homogeneous groups that rarely interacted with outsiders. These observations led Flack and her colleagues to hypothesize that the presence of police allowed for a healthier and denser network, where members had more and friendlier contacts with larger numbers of their fellows.
This kind of knockout experiment, which revealed that some individuals in a network are especially valuable to its structure, suggests that an understanding of animal social networks may be important to conservation biology. Take the case of killer whales (Orcinus orca). Individual juvenile females and clusters of related females appear to be key hubs for the transmission of information about foraging opportunities and other aspects of life in the sea. Anything people do that disturbs such individuals or group information hubs—from hunting to polluting the oceans to constructing barriers that impede the ability of whales to swim freely in their environment—might severely disrupt the killer whale social network and weaken the prospects for survival of the entire group. This understanding could, at the very least, inform policies to minimize the impact of our actions on these amazing creatures.
The bird song-and-dance network
Wild populations of birds in their natural habitats have also been the subject of social network analyses. One such species is Central America's long-tailed manakin (Chiroxiphia linearis). Males are strikingly handsome, distinguished by their indigo feathers, red cap and, as the name indicates, long, thin tails. Find the right pair of males together on a perch, and a bird-watcher can witness one sweet song-and-dance routine. Female manakins watch, too, assessing these performances when selecting mates. For males, the chance to perform matters—a lot. Unfortunately for them, competition for the chance at a spot in a duet is high and often quite aggressive.
David McDonald of the University of Wyoming spent more than 10 years in Costa Rica, totaling 9,288 hours observing the birds. He discovered, using social network analysis techniques, that it is males with a high degree of connectedness during their early lives that get the privilege of performing at this avian open-mic night.
Like any dance battle, it is all quite complicated, but what happens is something like this: Clusters of eight to 15 males spend their time in “perch zones,” areas that contain one to several perches, where the birds will ultimately perform. Any male in a cluster can practice singing and dancing on a perch outside the breeding season (late February to early September) or even during the breeding season as long as females are nowhere around. But with females present during the breeding season, only the two highest-ranking males—labeled alpha and beta—can sing and dance on perches. The competing performers actually become a team to aggressively dismiss all other males from the area.
The alpha male wins almost all matings in a perch zone. The payoff for the beta male is succession to the coveted top position once the reigning alpha male dies. This system creates a huge benefit for both alpha and beta males, one that all males want but that few get.
As young males mature between ages one and six, they often move between perch zones, establishing relationships with many other males. The average age of a successful breeding male is 10 years, meaning that any given male has many other males in his social network as he matures. In his nearly 10,000 hours of fieldwork, McDonald tracked which males interacted with one another each year for more than 10 years. He built a social network map from his data to see if the structure of the network would reveal which males ended up as successful duet singers.
His network metrics took account of both short pathways that connected one individual directly to another and indirect pathways that could include interactions between birds several links removed from the first individual. (“I don't know Bert personally, but I know Kermit, who knows Ernie, who knows Bert.”) McDonald ultimately determined that centrality was the secret: central males were much more likely than less well-connected ones to rise up in the breeding hierarchy, sometimes achieving the alpha and beta statuses that would allow them onstage to sing and dance their way into females' hearts.
It is important to note at this point that this kind of research identifies network structures and associates such structures with observed behavior. A direct causal connection between the structure and behavior is assumed—but not proved. Conceivably, alpha and beta males, rather than gaining power because of their many connections, could have gained many connections because of some features that made them popular among their peers.
The dolphin fishing network
As we previously mentioned, many tools of social network theory were imported from the social sciences. It is unsurprising, then, that some of the first subjects of detailed nonhuman social network analysis were bottlenose dolphins, already recognized as big-brained, intelligent, highly social animals, like us (on our good days).
In the late 1990s then graduate student David Lusseau fell in love with the common bottlenose dolphins (Tursiops truncatus) of Doubtful Sound, a gorgeous fjord in southern New Zealand, more than 200 miles west of the University of Otago, where Lusseau, now at the University of Aberdeen in Scotland, was working on his doctoral dissertation. For seven years Lusseau tracked these beautiful animals. One of his tools was photography, which helped him methodically identify the natural markings on each of the 64 dolphins in Doubtful Sound and to better follow those individuals.
After observing more than 1,000 pods of various sizes that contained subsets of these 64 animals, Lusseau determined that the dolphins were part of one large social network linking virtually all of them. Furthermore, he found that individual dolphins clearly preferred the company of only some other specific group members. But he could not put his finger on why. What did dolphin networking accomplish, and what manner of information or benefits might be shared among associates?
To investigate those questions further, Lusseau joined forces with Paulo C. Simões-Lopes of the Laboratory of Aquatic Mammals at the Federal University of Santa Catarina in Brazil. They studied a population of 55 bottlenose dolphins that were living on the other side of the planet and engaging in a unique behavior that Simões-Lopes had identified a few years earlier—a mutually beneficial interaction with the local artisan fishers (Homo sapiens).
Each spring, from April to June, fishers in the Laguna region of Brazil use a technique introduced to the area by settlers from Portugal's Azores more than 200 years ago. They cast long nets into the water to catch schools of mullet (Mugil platanus) migrating from the cooler waters off of Argentina. In recent years they have received help: some—but only some—of the bottlenose dolphins in the lagoons actually herd schools of mullet toward the fishers. At the right time, the dolphins slap the water with their head or tail. The slaps tell their human partners when and where to cast their nets. The upshot of this remarkable interplay is that both species of mammals catch more fish than they otherwise would.
Lusseau's previous experience made him consider social network analysis as a way to scrutinize the details of this rather incredible behavior. From September 2007 to September 2009, Lusseau, Simões-Lopes and some of their colleagues went out on boats in the lagoon system, dolphin photographs in hand, and gathered data on which dolphins were swimming together. The research team was able to collect reliable data on 35 of the 55 individuals in this population. Even this incomplete data set made it clear that these dolphins had established a highly structured social network.
A statistical analysis found that the Laguna dolphins could be subdivided into three cliques within which individuals spent most of their time. Although all the dolphins in any one clique had some tenuous interactions with the dolphins of the other cliques, animals within these cliques tended to swim together and interact mostly with the other individuals in their clique. Such close-knit associations could serve to facilitate information transmission among members.
Clique 1 consisted of 15 dolphins, every one of which cooperated with the local fishers. This clique was highly interconnected, with all its dolphins often associating with one another, both during the autumn mullet fishing season and during the remainder of the year. Information flow was therefore easy. Not surprisingly, clique 1 benefited from its relationship with the fishers, whereas the other cliques lost out on this opportunity.
Cliques 2 and 3 differed greatly from clique 1. None of the dozen dolphins that formed clique 2 cooperated with the fishers. And although clique 2 dolphins were often found together, both in and out of fishing season, their social relationships were weaker than those seen for individuals in clique 1.
Of the eight clique 3 animals, seven did not cooperate with fishers—but one dolphin, labeled “20”—did. And of all the dolphins in the Laguna population, dolphin 20 spent the most time interacting across cliques. It appears that dolphin 20 acted as a liaison between its clique and cooperative clique 1. Determining the influence of such liaisons in highly complex networks should be a fertile area for future study. The findings already suggest, though, that having a tight network, as clique 1 did, can help animals overcome challenges that individuals cannot solve alone—in this case, devising a way to communicate effectively with members of another species: human fishers.
The researchers do not yet know whether some key clique 1 individuals—perhaps older, experienced dolphins—teach other group members how to cooperate with the fishers. But given that teaching has been found for other complex feeding behaviors observed in dolphins, it would not be surprising to find similar instruction going on here. Indeed, such socially learned traditions form the basis of animal culture and are strongly facilitated by social networks.
Attitudes toward animals have evolved greatly since the early conception of them as robots mindlessly carrying out genetic programs. Ethologists now know that many animals are much smarter, more behaviorally flexible and better able to learn than the pioneers of the field could have dreamed. We anticipate that more studies of social networks, along with greater exposure to those studies, will further change the way that people think about animals. Hardly preprogrammed automatons, many nonhuman creatures spend their lives, as we do, within a complex social milieu—in networks where both direct and indirect interactions with other individuals drive so much of what matters to survival and success.