The Social Lives of Asian Elephants

Kanthi (far left) and Kamala (far right) were the most inseparable pair of elephants we saw during the study. Members of their social group, the K unit, were often together whenever they were seen.  Yet not all social units were so tightly knit, with individuals being scattered into small groups quite far apart.

Male and female Asian elephants form distinct parallel societies in which adult females and calves move together and form visible groups whereas adult males are typically more solitary.  For many years there have been two somewhat conflicting characterizations of female Asian elephant society.  The classic view, popularly held, is that Asian elephants form very tightly-bonded families centered around older adult females known as matriarchs.  This view is adapted wholesale from the many excellent long-term studies of African savannah elephants [1-3], which do exhibit this type of social organization.

The second view, held by scientists more recently, is that in fact Asian elephants have very small social groups, typically containing 3-5 individuals, and that even family members do not seem to spend a lot of time with one another. This second view was based on field studies of Asian elephant behavior as well as genetic analyses, but with relatively few individuals repeatedly observed, and conducted over very limited stretches of time [4-5].  Until now we have not been able to resolve this seeming disparity in our understanding of the social behavior of the largest herbivore in Asia, a species whose behavior in the wild has remained enigmatic despite its long historical association with people.

In a new study we’ve just published in BMC Ecology analyzing twenty months of continuous observations at Uda Walawe National Park, we show that the social behavior of female Asian elephants is something in-between these two extremes [6].  They form highly dynamic social units in which the relatedness is not always clear. Individuals change their immediate companions quite frequently – on the scale of hours or days – but maintain a very stable pool of associates over longer timescales such as seasons or years.  Often the long-term pool of companions is a great deal larger than the herd an individual might be seen with on any given day.  A social unit in fact frequently contains upwards of 10 adult females and sometimes many as 50!  In other words, a ‘herd’ of elephants one sees at any given time is just a small splinter group from the much larger social unit. This type of social organization is aptly named ‘fission-fusion,’ denoting that social units merge and split, and that affiliations are not static.  Other species that exhibit similar types of social organization include killer whales, dolphins, chimpanzees, and spider monkeys.

The study takes apart elephant society at three different levels of organization, with the intriguing finding that different patterns emerge at each.  At the first level we simply look at the relationships among pairs of individuals over time.  We find that the vast majority of pairs associate very strongly only for a single season.   A few pairs of individuals associate more strongly in dry seasons than wet seasons, showing a cyclic strengthening and weakening of their over the two-year period.  Individuals also vary a lot in their social strategies – some have a few preferred companions which they maintain faithfully over the long term, whereas others have many companions that they shuffle from season to season.  This makes a lot of intuitive sense if you think about it: the more companions you have, the less time you can spend with any single individual unless you maintain a few core companions who are your best friends.

This is what many Asian elephants do, it turns out.  At the second level, we started to look at the social networks of Asian elephants (no, not Facebook).  We first took a look at the most basic social network one can construct: the ‘ego-network’.  If the ego is you, your ‘ego-network’ is the set of all individuals you know directly.  In the case of elephants, it was the set of all individuals an adult female was seen with in any particular season.  We looked at how very simple properties of these networks changed over time – properties like how many friends an individual had, and the proportion of these friends who also knew one another, etc.  We found that each of these properties changed from season to season with no apparent similarities between wet seasons or dry seasons.  But the ego-networks also showed that there were small sets of animals who were consistently seen together over multiple seasons (see diagram below).

Lastly, we zoomed our view out to look at the entire population – around 286 adult females in total.  We found that the social networks for this large set of individuals were extremely complex and that they changed from season to season – in fact, no two seasons were alike at all.  Despite these changes, the patterns that emerged were far from random.  What we saw was that there were distinctive clusters – large, diffuse social units that persisted over the entire study period and longer.

 

Figure 5 Social networks by season. Nodes represent adult females and the thickness of edges corresponds to the SRI value. Isolates appear in the upper left corner of each network. Each row corresponds to a season and each column to a type of network. ‘Ego’ shows the egonetworks of selected subjects, who are indicated by black circles with colored borders. Node colors correspond to each Ego’s Girvan-Newman cluster assignments in T1. These colors are maintained through all seasons to allow comparison across seasons (actual cluster designations in other seasons are not shown). In subsequent seasons, individuals colored gray are those who did not appear as Ego’s companions in any preceding season covered by this study, although they may have associated prior to January 2007. Labeled nodes indicate those who associated with the subject in nearly every season. By the fifth season, networks clearly consist primarily of individuals who previously associated with the subject, even if not all were present in every season. ‘Residents’ shows all residents. One can track the coherence of each cluster over time. Some clusters from T1 maintain their integrity (e.g. brown) whereas others do not (e.g. light blue); associations among clusters also change over time, most notably the large pink and dark green clusters. Clusters are connected via just a few bridging individuals. ‘Population’ shows the full social network for each season with sample sizes reported in Table 2. Social networks constructed from real data have distinct structure, whereas those constructed from randomized data do not (included with paper, but not shown here).

When elephants separate from one another, they can find each other again over relatively long distances using sound and scent.  Because Asian elephants do change their social preferences so frequently, it is not clear what role matriarchs might play in this society and it’s an active area of our current research.  Additionally, we don’t know why different individuals employ different social strategies – is it the case that some females prefer one another because they are very closely related?  Or simply because they have calves of similar ages?  Is there a difference between individuals that are resident in the park and those that we see only periodically, which hints that a part of their homerange also lies elsewhere?  And what about other Asian elephant populations, in other habitats – in India for example, where some elephant populations occupy very dry environments and others occupy areas more similar to Uda Walawe in terms of rainfall?  And what might account for the differences between Asian and African elephants? All of these are questions left wide open for future research.

See full article here.

_________

Press:

A radio interview on Voice of America related to this article, with accompanying pictures and elephant sounds, can be found here.  Correction: the speaker refers to elephant ‘squeaks’ but the audio recordings are of ‘squeals’.

Science Daily – Social Networking Elephants Never Forget

ScienceNOW – Asian Elephants Are Social Networkers

Livescience/MSNBC – Social Networks Rule Among Asian Elephants

New York Times – Tracing Social Networks Of The Asian Elephant

Asian Scientist – Social-networking Asian Elephant Style

Zeit Online – Asian Elephants Also Have Social Networks (German language newpaper)

——————————————————————————————

[1] Douglas-Hamilton, I. (1972). On the ecology and behaviour of the African elephant Oxford University

[2] Moss, C., & Poole, J. (1983). Relationships and social structure of African elephants Primate Social Relationships: An Integrated Approach, 315-325

[3] WITTEMYER, G., DOUGLASHAMILTON, I., & GETZ, W. (2005). The socioecology of elephants: analysis of the processes creating multitiered social structures Animal Behaviour, 69 (6), 1357-1371 DOI: 10.1016/j.anbehav.2004.08.018

[4] Fernando, P., & Lande, R. (2000). Molecular genetic and behavioral analysis of social organization in the Asian elephant ( Elephas maximus) Behavioral Ecology and Sociobiology, 48 (1), 84-91 DOI: 10.1007/s002650000218

[5] Vidya, T., & Sukumar, R. (2005). Social organization of the Asian elephant (Elephas maximus) in southern India inferred from microsatellite DNA Journal of Ethology, 23 (2), 205-210 DOI: 10.1007/s10164-005-0144-8

[6] de Silva S, Ranjeewa AD, & Kryazhimskiy S (2011). The dynamics of social networks among female Asian elephants. BMC ecology, 11 PMID: 21794147 Full text and PDF.