These optical illusions are why butterflies are so hard to catch

Why butterflies are hard to catch

0:00Welcome back to the Nature Podcast. This time, how butterflies can trick you into thinking they're flying one way when they're really going another. I'm Nick Pertridge-How.

0:45Butterflies are hard to catch, perhaps uniquely so, and new research in Nature proposes that this is down to their patterns, and how they confuse predators. Now, you may think that butterflies are pretty and colourful to ward off predators. Don't eat me, their stripes say, I'm poisonous. Or that they're useful to attract mates. A dull butterfly may not stand a chance against a charming, colourful competitor. And you'd be right. But that doesn't explain everything.

1:19For instance, butterflies in Europe aren't poisonous, but still have the bright patterns that suggest they are. It seems these patterns may also be doing something more. And the new study reveals something new. These patterns are sending predators the wrong way with a confusing motion cue. Curious about these confusing colourations, I made a trip to Cornwall to meet the authors of the New Nature paper.

2:02Hi, Jolion. Lovely to meet you. Nice to meet you. That's Jolion Trushanko, one of the authors of the new study, as he welcomed me into his office. He is a visual ecologist who studies how animals interact with the world they see. And while there are a lot of ways that animals use visual cues to avoid predators, camouflage being the obvious example, he told me that butterflies may be uniquely good at this. The really, really telling, interesting thing is that there are no birds that have evolved to really

2:34specialise on catching butterflies in flight. So lots of things will try and eat moths and other insects. But butterflies, despite being the most obvious thing, they're so visually salient, you know, there's nothing much that really specialises on eating on the wing. When you see a bird try and catch a butterfly, they sort of just try and chase around it. They can never actually grab it. So this is an interesting tell that there's something weird going on with butterflies. It's rare, but some birds do actually catch butterflies, but they do it when they've landed. When a butterfly is

3:04in flight or on the wing, there's no known predators that specialise in catching them. In fact, I even got to see quite how hard it is to catch butterflies in person. Oh no! Oh, sod!

3:20This is George Hancock, one of Julian's colleagues and another of the authors on the new paper, trying and failing to catch a butterfly. Now, in fairness to him, he did quite expertly catch another two butterflies later on, but those had landed. Butterflies in flight seem particularly hard to catch, even with a net. Sir Julian and George were curious why this was, and they thought

The barber pole illusion

3:45that they could look to advanced camera technology to figure it out. These days, high-speed footage that you can slow down to a crawl is fairly commonplace. Sir Julian tried to find some of this footage and run it through a vision model, which replicates how birds' eyes see to understand what was the cause of the butterfly's unique weirdness. I just went on YouTube and took a high-speed video of a butterfly taking off. It was a BBC Natural History video, and I just thought, right, I'll take this, put it through the model, see what happens. And I thought I'd made a big mistake. It was colour

4:19coded, so green was upwards motion and purple was coded as downwards motion. And this butterfly was taking off, and it was just loads of this purple, this downwards motion was really powerful. And this is before I'd really conceptually thought about how this visual illusion might work. And I thought, initially, I've made an error, my code's gone wrong, so I did lots of double checking, and then, no, it was true. Despite the butterfly clearly moving upwards as it took off, Julian's model was identifying downwards motion. This is really this eureka moment of realising, wow, the way the stripes are

4:53angling in the different points in the wing beat cycle. They are turning into what we call a barber pole effect. They are creating sideways and downwards motion from a wing that's actually moving up. The barber pole effect gets its name from, you guessed it, barber's poles. These are the often red and white rotating poles that you see outside of barber's. They're made of a cylinder painted with a diagonal spiral stripe. This spins sideways, but looks as though it's moving up or down. This is

5:26because the information about the movement is ambiguous to our brains. We don't know what's happening to the stripes behind the pole. But because the stripes are diagonal, our brains interpret the movement as going up or down, as if they're twisting around the pole and then disappearing off the end. In essence, the stripey pattern misleads the eyes. Julian thought that a similar thing was happening with the butterflies, but he needed more than a single video nicked off YouTube to make sure.

5:56So he, and mostly George to be honest, caught some butterflies and then recorded them taking off with their own high-speed camera. And we took those videos, put them through this bird motion vision model, and then a really fun thing that we could do is, in each frame of the video, we could take away the butterfly's patterning. So we could simulate, here's the butterfly taking off with its real wing patterns, and then we could, in every frame, get rid of those patterns, make them an average grey or white or black. So simulating the same butterfly moving, same flight, same exact dynamics, but taking away the patterns.

6:31And then we can put both of those through the model and see how much patterns interfere with motion detection. And that was really cool because that showed that the presence of patterns had a really powerful effect at making it look like they're flying in the wrong direction. As the butterfly's wings move up and down, the stripes and patterns create illusions that make it look like the butterfly is moving slower than it is. That would mean a would-be predator would end up attacking behind the butterfly, missing all the important bits and allowing the insect to escape.

7:04But that's just what the model implied. The researchers wanted to see how this would work when an actual predator is trying to catch a butterfly. Unfortunately, such an experiment would be really difficult to do with birds. But George came up with a solution. He made a game that humans could play. In the game, they could see how much the butterfly's patterns interfered with the humans' abilities to net them. So in this game, your goal is to try and catch a butterfly. You're going to try and catch it as fast and accurately as you can. There is going to be a tutorial where it teaches you how to play.

7:36This is George explaining the game to a player. They have to catch a virtual butterfly on a touchscreen. George and the team can monitor how successful they are at catching it, as well as exactly where their missed attempts are landing. What we find by getting people to play this game is that the more motion confusing you are, the more attacks are directed towards the back of a butterfly, the more likely you are to hit less vulnerable parts of the wings. So the hind wings of many butterfly species, especially those such as swallowtails, which have a long tail, are evolved to break off. So if a predator just catches

8:11a butterfly but only catches that part, it's not going to come away with a meal and the butterfly is going to be able to continue to flee with its life. The human butterfly predators were also quite bad at catching butterflies. I can personally attest to this, having played the game myself. On average, people caught only 40% of butterflies, if only my own percentage had been that high. But as George says, even when the people hit them, they tended not to hit anything vital, being directed to hit towards the back of the wings. Survival, it seems, is the name of the game.

8:43Losing a bit of the hind wing isn't a big deal to a lot of butterflies, but surviving to mate and bring forth the next generation is. Which is what the team wanted to look at in their next experiment.

Simulating evolution in computers

8:56It seemed that butterflies had confusing motion cues. They could mislead predators, or at least humans playing a game, into missing them or hitting non-vital parts. But if this is the case, then you'd expect it to be an evolved trait. Rather than waiting to see evolution play out over the next few million years, the team used a computer simulation to fast-forward through generations of natural selection. And what we found was that if you select for our motion confusion metrics, you get butterflies that

9:28eerily resemble real butterflies that we see in the natural world. So if you select for a metric that we called forwards confusion, you get the same kind of stripe patterns you see all across swallowtail butterflies and various different nymph valids. And they've all settled on having these stripes. If you've ever seen a tiger swallowtail butterfly, or a scarce swallowtail butterfly, they have stripes in almost exactly the same positions on their wings. And we think this is because they have evolved in order to try and redirect attacks towards the back of them onto this tail, which can break off.

10:02George and the team think that this observation really validates their study. It's another bit of data along with the high speed camera footage, the game and the real world observations that point towards the butterflies misleading their would-be predators by creating confusing cues with their patterns. But one thing that really annoyed me was whites. So the white butterfly, this wasn't explained. It's got no, it's an immaculate white butterfly often, or it sometimes has little black spots on its wingtips. This wasn't really being explained by our model. And so another thing we did in

10:32the genetic algorithm model, this fully computer-based thing, is we flipped it. So instead of making the butterfly look like it's flying too slowly, too far backwards. What if we try and make it look like it's flying faster than it really is, so that the bird's attacks might be deflected too far forwards. And then they all turn immaculate white, almost instantly, sometimes with the little black spots on the on the wingtips too. So this is a really nice, it's an elegant way I feel of showing that this is, it's a strategy where there's lots of different ways of achieving this. And this fits in really nicely

11:03with how we think evolution should work for kind of creating lots of diversity, lots of different strategies for evading being eaten by birds. Julian and the team think that the method of avoiding predators that they've explained could be the most successful visual defence strategy in moving animals. By manipulating their predators with motion cues, the butterflies are able to avoid being caught. This probably is also aided by the butterflies' rather chaotic flying style. They change direction seemingly at random and can barrel

11:35through the air, flapping madly at 70 centimetres a second, about a mile and a half per hour. Pretty fast when you're only a few centimetres long yourself. And the team are currently investigating how butterflies' flying style affects this. One big caveat here though is that there's a lot we don't understand about how motion vision works. We also don't really know a lot about how birds see, which may have implications for the game the team created. While the game simulated how we think birds

12:06see to the humans playing it, it'll be better to have actual birds try it out. Something that's very, very difficult to do. I don't for instance envy the researchers who need to corral flying birds in their lab to hit butterflies. Ultimately though, Julian and George think that it's likely that many other animals influence how their predators see them using similar motion tricks. It seems to be a widely distributed trait amongst butterflies according to their study, and there's no reason to think that other animals don't also do it. Zebra stripes for instance have puzzled researchers for

12:40hundreds of years, so this work may help uncover the real reason why they have their stripes. Perhaps they too set off confusing cues that mislead their predators. So in the end, Julian and George hope that this work will inspire investigations of other animals. It's really exciting when you think that there are so many animals that need to be protected from being eaten while in motion. If you think about fish and birds, many of these animals have really high contrast markings on their

13:11wings or on their tails, and a really important thing here that we've kind of realised is this is due to biological motion. So real animals have limbs that at some points are moving in the opposite direction to where they're going. So on the butterfly their wings are moving downwards at one point, so you maximise the motion on that downwards moving limb, and then hide the motion, fold that wing up and hide it on the way up. And there are lots of animals that can do this. As a fish is escaping its predators, the spots on its tail will at one point move in the opposite direction to the rest of the fish and where

13:43it's going to be in 100 milliseconds time. And so this is potentially super common in nature. And yeah, we've now hopefully got the tools to start exploring this. That was Julian Trushanko from the University of Exeter here in the UK. You also heard from George Hancock, also from the University of Exeter. For more on that story, check out the show notes for some links. Right now though,

Loud birds and iron age shipwrecks

14:06it's time for the Research Highlights with Dan Fox. Don't worry, you're not receiving an emergency alert. That is the call of the bare-throated bellbird, one of two birds in contention for the title of world's loudest bird. The other, the red-legged ceriema, sounds like this. We've turned it down a bit to save your eardrums,

14:39but trust us, at full volume, it is loud. These two avian ear splitters were identified by researchers investigating the role of evolutionary biology on bird communication. Being loud is vital for communication and this team set out to answer two questions about bird calls. To what extent can birds vary their vocal amplitude? And what are the evolutionary factors that drive this? Using laser rangefinders designed for golfers and sound level meters intended to ensure workplace

15:14safety, they were able to make field recordings of over 100 species of bird at locations in the US and Brazil. Analyzing the data, they found 58 decibels of variation in bird calls, from a tiny hummingbird singing at 64 decibels, up to the booming duo we heard earlier peaking at 120 decibels, louder than a chainsaw. They found that bigger birds tended to be louder, with wider beak openings and stronger muscles to compress and explosively expel air. But birds are often louder than significantly larger land

15:52animals. And so, the authors suggest, other mechanisms enabling their extreme vocalizations remain to be discovered. You can read that research in full in Evolution.

16:07In Iron Age Europe, around 2200 years ago, iron was, unsurprisingly, pretty important. But the scale of its production and trade has been unclear. Now, the discovery of a huge cache of iron bars suggests that iron was being shipped along the river Danube more than a tonne at a time. Researchers analysed 193 complete iron bars and hundreds of fragments recovered during gravel extraction along the Danube in northern Austria. The pieces represent at least 500 forged iron bars,

16:44making the find the largest known assemblage of such objects from Iron Age Europe. Radio carbon dating of three bars, together with analysis of similar artefacts found elsewhere, suggests that the cargo dates to the 3rd or 2nd century BC, and would have been enough metal to equip around 500 warriors with swords. The authors say the iron was probably being transported down the Danube when it was lost in a shipwreck, pointing to large-scale iron production in Central Europe during the late Iron Age. You can find that paper in antiquity.

17:19And that's all for now. If you want to stay in touch with us, you can. We're on email, podcast at nature.com, or you can find us on social media, at naturepodcast. We'll be back again soon with more from the world of science. Until then, I've been Nick Petra-Ciao, thanks for listening.