Olive baboon. Credit: SajjadF (CC BY-SA 3.0)
Olive baboon. Credit: SajjadF (CC BY-SA 3.0)

Baboon-Trackers Herald New Age of Animal Behaviour Research

Picture a troop of olive baboons, moving over the savannah. There’s around fifty of them, and they cover a lot of ground as they search for grass, seeds, insects, and other bits of food. They need to stick together so they don’t get eaten, but different animals might want to head in different directions at any one time. How do they coordinate their choices to preserve the sanctity of their group? As primate researcher Joan Silk says, “It’s hard enough to get two adults and two kids into the car at the same time let alone 50 baboons who can’t talk.”

It’s a fairly simple question—how do animals make decisions as a group?—but it’s also incredibly hard to answer, even for animals like baboons that have been actively studied since the 1950s. A field-worker can easily track a single baboon, or even a few. But she can’t track all 50 at once, let alone note down who’s getting up, who’s leading, and who’s following, over long periods of time.

She can, however, turn to technology. By fitting wild olive baboons with sophisticated GPS collars, which automatically record their movements, Margaret Crofoot from University of California, Davis had learned exactly how they make decisions about where to go and whom to follow. Some of her results are surprising, others are more intuitive; regardless, her study heralds a new age of zoology, in which scientists can analyse animal behaviours on a scale that was previously impossible.

For several years, Crofoot had been using radio-trackers on wild capuchin monkeys, to study how they fend off predators and fight other monkeys as a group. “I was getting increasingly interested in how individual decisions scale up to group behaviours,” she says, “but you can’t watch everyone at once.” Then, she heard about the collars.

GPS collars aren’t new but they often have poor temporal resolution. They might take a reading once every few hours, for example—good enough to track a migrating bird. Newer models are much more sensitive and can take readings once a second—good enough to track a running cheetah. “I thought, wow, if you could put these collars on primates, you could do all these studies on an intact social group, living in its natural habitat,” says Crofoot.

Olive baboons were an obvious choice. They’re well-studied, so their behaviours can be interpreted in the light of decades of knowledge. They live in open habitats, where GPS signals are strong and good. And unlike many other monkeys which scamper through trees, baboons live on the ground, making them easier to follow, trap, and collar.

View Images
A troop of olive baboons. Credit: Stig Nygaard CC-BY-2.0

The hardest bit was actually finding the right troop. Many researchers have spent a lot of time in habituating baboons to their presence, and they were concerned that trapping the individuals would wreck that hard-earned tolerance. So Crofoot worked with a troop in Kenya’s Mpala Research Centre, which weren’t part in any other studies. Over a week, her team caught and collared 25 of the individuals.

Once the data were in, team members Ariana Strandburg-Peshkin and Damien Farine worked together to interpret it. They wrote a programme that would automatically identify events where, say, one baboon walked off and others followed, or when one walked off and then returned to the same spot. The video below shows 25 minutes of movements, sped up 25 times.

Surprisingly, they found that a baboon’s rank in the pecking order didn’t affect its odds of being followed. Rank matters a lot in baboon society, and affects how much sex, food, and support each individual gets. When making foraging decisions, the dominant males wield a despotic hold over the rest of the group, enforcing choices even when they’re the wrong ones.

But when it comes to more mundane decisions like “Where should we go?”, their tyrannical sway isn’t evident. The data revealed that the troop members didn’t weigh the movements of dominants any more heavily than those of subordinates. Age and sex didn’t matter either. “It’s a little surprising that dominants aren’t using their social power to drive group decision in ways that are beneficial to them,” says Crofoot. “It seems that on a day-to-day level, most decisions are made more democratically.”

The team also showed that what happens if baboons have differences of opinion. If two baboons move off in different directions, but at the same time, the angle between their paths determines what the others do. If that angle is small—say, less than 90 degrees—other baboons will split the difference and head off in the average direction. If the angle is big, the followers make a choice, and trail one initiator or the other.

That’s rather astonishing—not because of the rule, but because it exactly follows what Iain Couzin from Princeton University predicted a decade earlier, using just mathematical theory and computer simulations. Couzin, a leader in the field of collective behaviour and a collaborator in Crofoot’s study, modelled the movements of animal groups, creating digital swarms in which individuals were all the same and had no special relationships. By contrast, identities and relationships are paramount in baboon society. And still, “the simple model predicted the behaviour of the very complex social group,” says Crofoot. “We were really struck by just how closely the patterns matched.”

“It’s a great study, and really innovative,” says Joan Silk from the University of California, Los Angeles. “We’ve been perplexed for many, many years about how animals in groups figure out where to go next, and how the process of group movement is coordinated and negotiated. It’s been impossible to quantify the movement patterns of multiple individuals at the same time. But now, we can do this.”

This is just the beginning. The team have uploaded their gargantuan pile of data to Movebank—a free, online database of animal tracking data—so that others can have a play. They also have plans for more analyses of their own.

For example, Strandburg-Peshkin says that individuals baboons do vary in their odds of being followed—it’s just that this variation doesn’t correlate with anything obvious like sex, age, or dominance. What then? What makes one animal a likelier leader and another not? Is it something about where they sit in the group, in physical space rather than social hierarchy? For that matter, what determines where individuals sit, and how do their positions change as the group shifts and moves? And what happens when a stationary group decides to travel? “How do you overcome the inertia of whole bunch of baboons sitting around?” asks Crofoot.

Strandburg-Peshkin also notes that they only looked at data from one baboon troop, and she’d like to know if other groups—or other animals—show the same patterns. A decade ago, that would have been wishful thinking. Now, with improvements in GPS collars, flight trackers, and related tech, it’s a plausible expectation. We’re entering a new age in the study of animal behaviour.

But don’t neglect fieldwork! By casting their actual eyes and ears into the wild, researchers can understand the social structure and behavioural quirks of their animals—information that no collars can provide. The trick will be to marry that hard-won, old-fashioned knowledge with the world of big data. “That would be the dream,” says Strandburg-Peshkin.

“Most behavioural research on wild baboons requires years of study and yields very little data,” says Andrew King from Swansea University, who has done more work on baboons than most. “This study is an excellent example of how we can generate lots of data very quickly to complement long-term observational studies.”

“We are using similar tracking collars to understand baboon behaviour and ecology in Namibia and South Africa,” he adds. “The next few years will be a lot of fun.”

Reference: Strandburg-Peshkin, Farine, Couzin & Crofoot. 2015. Shared decision-making drives collective movement in wild baboons. Science http://dx.doi.org/10.1126/science.aaa5099