Social Cognition in a Non-Social Reptile? Gaze-Following in Red-Footed Tortoises

redfooted tortoise.jpg

Have you ever been at a party with lots of people chatting away, when for some unexplainable reason you felt compelled to turn and look at the front door of your friend's house...and just as you were looking, someone was just coming in from outside and closing the door? You couldn't have heard the door open since there was so much noise already inside - more likely you noticed that other people were looking at the front door. All of this probably happened without any explicit intention or awareness. If several others are all directing their attention at a specific point in space, there might be something important there. We're naturally aware of where others are looking. And so are lots of other animals.

ResearchBlogging.orgGaze-following is the ability of an animal to orient its gaze to match that of another animal, and though this ability has been observed in mammals and birds, the phylogeny of gaze-following is still uncertain. Lots of evidence has been found for gaze-following in humans and non-human primates, and in dogs and dolphins, with somewhat more ambiguous evidence in other mammals and birds. But gaze-sensitivity - the ability of an animal to avoid the gaze of another animal - seems to be somewhat more common in the animal kingdom, having been observed in mammals and birds, and some reptiles and fish. Gaze-sensitivity may have evolved as an anti-predator defense; a theory known as the "evil eye hypothesis" suggests that the awareness of the gaze direction of a predator would help an animal know when it was safe to move about or come out of a hiding spot. Gaze-following requires gaze-sensitivity; indeed, gaze-following develops in human children after gaze-sensitivity. It therefore follows that gaze-following is cognitively more complex than gaze-sensitivity.

Dr. Anna Wilkinson and her colleagues from the Cold-Blooded Cognition Lab wondered about the evolutionary history of gaze-sensitivity and gaze-following. One way to investigate this is by better accounting for gaze-sensitivity and gaze-following throughout the animal kingdom, beyond mammals and birds. Are these abilities also present in reptiles? If so, it could suggest that all amniotic species (birds, mammals, and reptiles) share them, and that it emerged quite a long time ago, in evolutionary terms.

The red-footed tortoise (Geochelone carbonaria) is a solitary, non-social species native to the tropical forests of Central and South America. Individuals of this species spend most of their lives alone, and only really interact for the purpose of mating. Despite this, it is possible that these animals have the basic building blocks of gaze-sensitivity, and in the proper circumstances, might be able to develop gaze-following with sufficient learning.


Eight captive-bred red-footed tortoises were socially housed for six months prior to this experiment. One tortoise, the demonstrator (the same individual was always used as demonstrator), was placed on one side of a tank, and a second tortoise, the observer, was placed on the opposite side of the tank. They were separated by transparent screens. Above, a small opaque partition separated the two sides of the tank. The investigators directed a small laser beam towards the opaque partition on the side of the demonstrator. Once the demonstrator noticed the light, she invariably looked up at it. The experimenters varied the color of the light to maintain her interest, such that she would not habituate to it. When the demonstrator looked up, would the observer direct his or her gaze up as well? If so, it would suggest that red-footed tortoises, despite their solitary existence, are sensitive to the gaze direction of their conspecifics.


Two control conditions were included as well. The "no lookup control" was identical to the experimental condition, except the experimenter did not direct a laser towards the opaque screen, such that the demonstrator tortoise would not look up (if the demonstrator happened to look up by chance, that trial was discarded). The "no demonstrator control" was identical to the experimental condition, except the demonstrator was absent - the experimenter still shined the light on the opaque screen, to rule out the possibility that the observer tortoises somehow noticed the light.


There was a clear difference between the conditions, with the observer tortoises looking up in the experimental condition significantly more than in either of the control conditions. This was the first study to demonstrate that reptiles are able to follow the gaze of conspecifics, suggesting that gaze following may occur more often in the animal kingdom than previously thought.

It is possible that the common ancestor of the three amniotic classes - birds, mammals, and reptiles - possessed the ability to co-orient and follow the gaze of others, rather than gaze-following having evolved two or three separate times. There was theoretically little selective pressure for such an ability to have emerged in this particular species, given their solitary lifestyle. Another possibility, however, is that gaze-sensitivity may be innate, and that gaze-following builds on this innate mechanism through associative learning. This could also explain the results of this experiment, as the tortoises had six months of social experiences prior to the beginning of the study. Ideally, this study could be repeated in red-footed tortoises that did not have significant social experience. It is possible that co-orienting is unique among reptiles to red-footed tortoises, though it is at least as likely that co-orienting behavior is common among the superorder chelonia - turtles, tortoises, and terrapins - among all reptiles, or potentially among all amniotes. At present, each explanation seems equally reasonable.

Wilkinson A, Mandl I, Bugnyar T, & Huber L (2010). Gaze following in the red-footed tortoise (Geochelone carbonaria). Animal Cognition, 13 (5), 765-9 PMID: 20411292

Red-footed tortoise image via University of Vienna Department of Cognitive Biology.

For more on cognition in these red-footed tortoises, see:
Cold-Blooded Cognition

For more on communication in reptiles, see:
Robot Lizard Push-ups
To Hear A Mockingbird: The Plight of the Iguana


More like this

Why can't researchers just observe these [redacted] animals without setting them up in some preposterous arrangement? It's fine to control an experiment, but when you isolate an animal from its environment and disrupt its natural way of living (and socializing), you are at best seeing a very small part of the big picture and at worst shooting yourself in the foot as someone who wants to know how the world works. We may as well stick to studying our pets if we are going to have wild tortoises jump through hoops. Fear neither parasite nor predator, and out into the field.

I'm dubious. Given the importance of the head-bob in reptile displays, this experiment seems likely to be heavily biased toward the gaze-following result.

Good catch, Chris (@1). The mating head-bob is actually horizontal, not vertical, and only occurs in males (and this co-orienting behavior was seen in both sexes). This is what the authors have to say:

...the behavior observed was not the result of innate courtship behavior. The male red-footed tortoiseâs courtship display involves highly distinct sideways movement of the head. This is very divergent from the looking up observed in this study. The two behaviors can not be confused. Furthermore, this courtship behavior is only seen in males (Auffenberg 1965), and the lookup behavior was observed in both sexes.

What an interesting discovery! Gotta wonder about the thought process of the turtles. "Hey, what's Molly looking at? Ooooh, pretty."

One minor quibble with the technical terms: because birds arose from a sauropsid ancestor (dinosaurs), birds are, themselves, reptiles, so there are only two amniote branches: synapsids and sauropsids.

@Jason, glad they did check on that! I'm curious whether there are non-mating head-bobs known in the species, for territorial display or similar. Still, I expect that the paper will contain data comparing interactions based on the genders of the interacting animals, which should control for that relatively well.

I keep a Geochelone sulcata tortoise named Charlie with a G. pardalis tortoise named Pardner. If I put salad in their feed dish and one tortoise sees it & starts moving towards it, the other who was facing away will turn around to see what his congeneric is going for. I don't know the genders of either of these tortoises. Both were hatchlings in 2000. They engage in shoving matches, which Charlie now usually wins since "he" has outgrown Pardner. I've never seen either engage in head bobbing behavior.

By darwinsdog (not verified) on 15 Nov 2010 #permalink

Gaze-following requires gaze-sensitivity; indeed, gaze-following develops in human children after gaze-sensitivity. It therefore follows that gaze-following is cognitively more complex than gaze-sensitivity.

More complex? Is that a definable term in this context?

And why the "follow"? -- gaze-following might be developmentally constructed from gaze-sensitivity, but that could simply be a function that the latter is more selected for than the former; it just precedes it chronologically, evolutionarily and developmentally -- but that doesn't imply any logical relationships between the two, just a contingent relationship.

It makes sense that gaze-following requires more computation than gaze-sensitivity -- after all, if you follow gaze, then it should be trivial to construct any other kind of response to gaze, including ones that take fewer steps from "sensing" to "response". It's just that the argument here for it seems poorly constructed. The "requires" seems to be a developmental requirement, when it's the logical entailment that's wanted.

The argument seems possible to put on firm grounding -- it just seems to be not firm here, and the danger lies in it being used as a template for other arguments where it doesn't "happen" to be right.

This post has a great intro which really makes it clear how this process doesn't require direct awareness (though it can still involve it). It seems more likely than not that the turtles don't "know" why they feel like looking there â but that doesn't make them much more instinct-driven than humans (in this particular case).

I do agree with frog, though. (And not just because I'm biologically compelled to imitate amphibians!)

Given that the tortoises have been kept together for six months, a conditioned behavior as a result of competition for food could be an explanation for the results. To follow a tortoise looking to something - presumably food being put in the tank - would be a behaviour rewarded with getting to eat earlier than if you don't follow the gaze of the first tortoise finding the food. This would be very different from social gaze following.

Both male and female red-footed tortoises, adult and juvenile, display face-to-face head-bobbing activity.

and they also are highly social and have been known to congregate (get together and hang out)

By Daniel L. O'Malley (not verified) on 22 Nov 2010 #permalink

I'm curious whether there are non-mating head-bobs known in the species, for territorial display or similar. Still, I expect that the paper will contain data comparing interactions based on the genders of the interacting animals, which should control for that relatively well.