29.Bats with maps?

Bats suffer from big PR problems. They are traditionally associated with all kinds of bad things - from vampires and the devil, to rabies and COVID-19. And many people think they’re ugly too, though I think this Egyptian fruit bat is actually very cute.

But one thing is beyond doubt: bats are very clever.

Many of them fly by night and have astonishing powers of echolocation, while others are day-flyers that rely on their excellent eyesight. Some bats migrate over large distances. And they’re all very good at navigation.

A new study from Lee Harten et al. at Tel Aviv University sheds light on the Egyptian fruit bat’s impressive navigational skills - and in particular, whether they make use of a ‘cognitive map’.

As I explain in Incredible Journeys, the cognitive map has been the Holy Grail of animal navigation studies ever since it was first proposed back in the 1940s by the great Berkeley psychologist, Edward Tolman.

Because rats sometimes use shortcuts when searching for food rewards in mazes - entirely novel routes that they have never used before - Tolman thought they might be making use some of some kind of ‘cognitive map’ on which they stored information about the layout of their surroundings.

This was a very controversial idea because rats were then thought capable only of learning fixed routes based on trial-and-error (otherwise known as ‘stimulus-response’ or ‘S-R’ learning). Acquiring map-like knowledge couldn’t easily be explained in S-R terms, so it seemed impossible - according to the prevailing behaviourist orthodoxy.

Tolman was attacked on all sides and his opponents came up with all sorts of alternative explanations for the behaviour of the rats - some of which were pretty bizarre. But when cognitive neuroscience took off in the 1960s, it became possible to monitor what was going on in the brains of living experimental animals. And many discoveries made since the 1970s have born out Tolman’s hypothesis.

To cut a long and complicated story short, there’s now abundant evidence that various mammals - including rats, mice and human beings - have brain circuitry that allows them to form cognitive maps of their surroundings, though there’s still plenty of room for debate about how these work in practice.

What is less clear is whether any other animals have the same abilities, though there are some scientists who would make that claim - even for insects like honey bees.

Bats of course are mammals and it would be quite surprising if they were any less gifted navigationally than, say, rats or mice. After all, they navigate over much larger distances - and they are obliged to do so in three dimensions!

The new study is interesting because it provides pretty solid evidence for the first time that fruit bats really do use cognitive maps.

Using GPS trackers, Harten et al. mapped every single journey taken by 22 young bats (‘pups’) starting with their first excursion outside the nest and continuing every night thereafter for five months.

The pups gradually extended their home ranges until these reached a mean area of about 60 square km. Sometimes they undertook exploratory flights which took them beyond into unfamiliar territory. When they detected new sources of food, they would return to them later on to feed.

The striking finding was that the pups - like Tolman’s rats - often performed impressive shortcuts. These were defined as routes in which at least 50% of the animal’s journey was novel (i.e. the pup passed no closer th100m to any location it had previously visited).

You may wonder whether these shortcuts were just chance events. Well, they certainly looked intentional.

The paths the pups followed were almost as straight as their regular ‘commuting’ routes to familiar sources of food, and they were much straighter than their exploratory flights. Moreover, the pups headed straight for their targets right from the outset of these flights. Some of the lengthier shortcuts (described by the researchers as ‘long-cuts’) occurred at the end of exploratory flights. These sometimes involved navigating for many kilometers over unfamiliar territory.

And out of the 246 short- and long-cuts that were observed, hardly any was predictable on the basis of a randomised (‘random-walk’) movement strategy.

Could the pups have been following a scent trail?

There is plenty of evidence that homing pigeons make use of olfactory cues to find their way back to their roosts - especially when taken to distant and unfamiliar locations (though this is still the subject of some debate). But Harten et al. found no correlation between wind direction and navigational straightness. On this basis they judged it unlikely that the bats were relying on olfaction, though they could not rule it out completely.

What about echolocation then?

The researchers note that the bats’ echolocation system has quite a limited range - for example, they can detect a large tree only when it is within 50 m.

Harten et al. also found no evidence that the bats were making the kind of errors that would be expected if they were relying on ‘path integration’ or DR to make their shortcuts.

So they concluded that the bats must have been relying primarily on their acute vision to perform these navigational feats.

But of course good eyesight alone isn’t enough.

Harten et al. believe that, before setting off, the bats look around them and use the spatial arrangements of distant landmarks (such as high-rise buildings) to judge where they are and the heading they need to follow to reach their goal.

And since the bats are not following the same routes again and again, they must have some way of storing the different locations of these landmarks (as well as their varied appearances) in a form that doesn’t depend on any fixed point of view.

So, it really does look as if they must have some kind of cognitive map, though it will be interesting to see how other experts react to the new study.

25.How hunting dogs find their way ‘home’ in a forest

Here’s a link to a fascinating recent piece of research that explores the homing abilities of hunting dogs with a very well-developed sense of smell - so-called ‘scent hounds’.

Katerina Benediktova and her colleagues put ‘action cams’ and GPS-tracking collars on 27 hunting dogs and let them roam freely around a forest. (NB: the dogs were not passively moved to a new location like the animals in the earlier experiments we’ve considered.) They then analysed how the dogs found their way back to their owners - across no fewer than 622 trials at 62 locations.

(Even if you don’t want to get into all the details, it’s worth watching this video clip - it’s really fun!)

The team predicted that the dogs would either follow their own scent trail back to their owner (‘tracking’) or they would take shortcuts - a strategy they called ‘scouting’.

And that’s exactly what they found.

In 399 cases (almost 60%) the dogs used a simple tracking strategy to retrace their outward route. But in 223 cases (33%) they homed by a novel route. Somehow they ‘scouted’ out a completely new path that would take them back to their owners more quickly and directly.

How did they do that?

The research team have pretty well ruled out the possibility that the dogs made use either of the sun or polarisation patterns in the sky to help them set a course. They also believe that the wind direction would in most cases have made it very difficult for the dogs to ‘scout’ their way back to their owners using only their acute sense of smell.

But they did discover something quite new.

The ‘scouting’ dogs typically performed a short north/south run just before setting off for home - a so-called ‘compass run’.

Benediktova and her team think that these ‘compass runs’ reflect the dogs’ ability to detect the Earth’s magnetic field. They suggest that the runs may help the dogs’ recalibrate their DR (or ‘path integration’ system). In other words, they may help the dogs eliminate errors that have accumulated in their estimates of their position relative to their starting points. So perhaps their homing system is based on DR.

While this might explain how the hunting dogs in this experiment homed successfully, it’s hard to see how the dogs in the earlier experiments could have used DR to find their way home after being passively displaced. How would a dog in a closed basket accurately keep track of its position after a circuitous journey to a location 89 km away?

While it now looks more likely than ever that a magnetic compass plays an important part in the amazing homing abilities of dogs, I wonder whether they also make use of some kind of ‘cognitive map’ that works in conjunction with their DR, olfactory and compass skills.

More research is needed to tease out these issues.

No doubt there’ll soon be more to say on this subject. I’ll try to keep you posted!

24.How guide dogs navigate…

In my last two posts I’ve mentioned the intriguing evidence that some kind of magnetic compass is involved in the remarkable homing abilities of dogs.

But a compass by itself only tells you which way you’re facing. It can’t tell you which way you should go - unless you already know the direction you need to take.

So how could a dog that wants to find its way home know which way to go ?

The most obvious possibility is that, like many other animals, dogs are good at DR (or ‘path integration’).

DR (as I’ve discussed in earlier posts) is the process of keeping track of your position by recording each and every course change as well as the distance covered between each one. By ‘integrating’ this information you can in principle set a course that will take you directly home.

An innovative recent experiment explored the navigational capabilities of 23 ‘expert’ guide dogs.

The details of the tests are quite complicated. So I’m just going to summarise the keys findings. But do click on the link above if you want to know more.

The dogs were first trained to respond to the command ‘find the van’ and then, with a blindfolded instructor as handler, they were given four different tasks, each of which was designed to explore a different aspect of navigation: path-retracing, homing, shortcutting and detouring.

The dogs were first taken along an indirect path from the designated ‘home’ point to the van - just once. This gave them an opportunity to learn the route. Perhaps surprisingly, only 30% of the dogs were able to retrace the path they had followed on the outward journey. 43% were able to retrace the return route (‘homing’).

But this rather poor performance is apparently typical: guide dogs normally need three or four trips before they can reliably reproduce a route.

Now comes the exciting part.

80% of the dogs were able to take successful shortcuts on their way home, and 87% were able to find their way when an obstacle was placed in their path and they had to make detours. Errors increased as the route got longer - not surprisingly.

So even if most of the dogs weren’t very good at learning a route they had followed just once, a large majority were able to keep track of where they needed in order to get ‘home’.

The most obvious explanation is that the dogs are performing DR in some way. In other words, they are able constantly to update their geographical relationship to their ‘home’ as they travel away from it.

A magnetic compass would plainly be very helpful in this process, but the dogs would also need to be able to measure distance quite accurately. Some kind of odometer would seem to be required - perhaps based on step-counting.

That seems plausible enough in the case of the guide dogs. The journeys they were making were quite short: only a few hundred meters. But is it really possible that the dogs in Müller’s experiments (see previous post) were able to find their way home, just by using DR?

Even with a magnetic compass to track course changes, that would be quite a feat - especially after being carried in a closed basket along a circuitous route to a point 89 km from home! How could his dogs have measured distance?

Could their sense of smell have played a part? Or is something else going on?

Watch this space!