All societies (human and animal) can be viewed as
networks of interconnected individuals, linked by social, spatial,
temporal, and other relationships. By studying network structure (the
links between individuals) we can derive unique insights into the
workings of the society, and better understand the behavioral
strategies that individuals use to enhance their success. Social
network analysis has been extensively developed and used for studying
human societies and human behavior. Animals are also social, and
researchers have applied social network analysis to animal societies
and behaviors as well. To date, surprisingly complex social networks
have been described in a wide range of species and taxa, including
primates, cetaceans, ungulates, rodents, birds, fish, and insects.
Although there are unique insights to be gained by studying animal
social networks, there are also unique problems. In particular, unlike
humans, animals will not tell us where they have been and who they
associate with, information that is key to studying a social network.
Direct observation methods are usually used study animal social
associations and movement patterns. However, in many cases observation
is not possible, especially when animals must be continuously monitored
for long periods of time, when animals behave cryptically (as they
often do), or when animals are out of sight of observers (as with
nocturnal, arboreal, or fossorial species).
For these reasons, new tools have been developed
to allow researchers to monitor social associations on a continuous
basis no matter where the animal subjects are. These wireless devices,
variously called encounter or proximity logging tags, are worn by all
study subjects and can detect and log the presence of other individuals
wearing similar tags. When two such tags come to within a preset range,
they exchange their unique ID codes and store the event as an
“encounter” in a log file stored in memory. Later, the encounter logs
are retrieved and analyzed to determine
the association network structure. A successful deployment of proximity
logging tags can generate hundreds of thousands of logs that provide
a complete record of all interactions among tagged individuals
over the lifespan of the tags.
Hypothetical encounter logs of associations
between 8 tagged subjects, and the resulting network graph. In the
graph, ties (red lines) indicate detected encounters, tie thickness
denotes encounter duration (longer encounters suggest stronger social
bonds). In network analysis parlance, animals 1 and 4 have more links
and therefore have greater centrality than the others in this group.
These individuals may have greater social influence and serve as
information hubs. Not shown in this figure, but potentially important,
are the timing of encounters, encounter location, how close individuals
were to each other, and behaviors that occurred during encounters. All
of these additional social parameters will be collected Encounternet.