Four red-footed tortoises have learned to use touchscreens to
carry out a series of tasks in a study of the reptiles’ spatial
There have been lots of studies into animal spatial cognition
involving mammals and birds, but much less research with reptiles.
Most of the research that has been carried out on reptiles has
involved sea turtle navigation.
Wilkinson from the University of Lincoln and colleagues sought
to explore spatial learning in the red-footed tortoise, a
land-dwelling creature native to Central and South America. The
tortoise has good colour vision and is food-motivated, making it
suitable for studying visual-based spatial learning.
Previous studies have shown that the tortoise is capable of
mastering an eight-arm radial maze, where it was required to
remember several different locations within a single trail. The
study seemed to show that a second tortoise could learn a path that
leads to a goal simply by observing the first demonstrator
tortoise. Perhaps more impressive is the fact that the second
tortoise didn’t simply learn the demonstrated route, but it was
able to apply the principles of the task even when the path to the
food was altered by introducing different turns. This suggests that
the tortoises are able to generalise knowledge across variations of
a previously learned task.
In this new study, the aim was to see how tortoises interacted
with a two-dimensional touchscreen display and compare that to a
traditional testing arena involving three-dimensional objects.
Four tortoises — Esme, Molly, Quinn and Emily — were put in
front of a touchscreen. A feeder hole was positioned in the centre
of the floor, 2cm away from the touchscreen. When tested, a correct
response resulted in the feeder producing a reward of a piece of
apple, cucumber or grape.
Developing the feeder, Wilkinson told Wired.co.uk, represented
the biggest challenge. “Automatic feeders generally use dried food
and tortoises like sticky fruit, so we had to develop a totally
novel way of delivering the reward. We used a large disc that
rotated after every correct choice, presenting a piece of fruit
under a feeding hole. The solution worked very well.”
In the experiments, the screen would first show a red triangle.
Once the red triangle was touched it would disappear and two blue
circles would appear on the screen. One of the circles would be
designated as the one to release the treat. If the tortoise chose
the correct one, it would be rewarded. If it chose the wrong one,
there would be a three-second timeout before the trial was
After the touchscreen training, the tortoises were tested in the
arena. In the arena there was a barrier blocking their view which,
when lifted away, revealed two blue bowls in front of the animal.
The aim was to see whether the tortoise continued to go towards the
“correct” blue target that it had learnt in the digital sphere.
After this test, the researchers reversed the reward side — so if
the animal had initially been rewarded for going to the left blue
circle or blue bowl, the team would now only reward the animal if
it went to the right. The tortoises were then sent back to the
touchscreen environment to see which blue circle they were
conditioned to peck at.
The results showed that the red-footed tortoises were capable of
learning to operate a touchscreen and there was some indication of
an ability to transfer learning from a touchscreen to a 3D test
arena, although no firm conclusions could be drawn as only two of
the tortoises showed this ability. Further work with a larger
sample of tortoises will be needed in order to firm up these