Steve Mirsky:    Welcome to Scientific American Science Talk, posted on May 16, 2015.  I'm Steve Mirsky.  On this episode—

Lee Dugatkin:    So what the dolphins do is they provide these very, very coordinated, specific cues to the fishermen to indicate that the fish that the fishermen are looking for are present.

Mirsky:      That's Lee Dugatkin.  He's an evolutionary biologist and behavioral ecologist at the University of Louisville.  He's the author of the book Principles of Animal Behavior and he's the co-author of an article in the June issue of Scientific American called "The Networked Animal," along with Matthew Hasenjager, a doctoral candidate in Dugatkin's lab.

I called Dugatkin to talk about the article.  He recorded his side of the conversation in his office, which is apparently right next to the Louisville Airport.

First, you just got back from Mongolia, which is not a place that most people in the U.S. have visited.  What were you doing there and what was it like?

Dugatkin:  Yeah, I was there for about eight days and I have a colleague at the National University of Mongolia.  And he had me come over.  I gave a couple of classic research talks on my own work, the kind of things that were going on in my laboratory.  But I was also there to do a two-day crash course on behavioral evolution.

         The Mongolians have a really strong program in ecology, but they don't really get trained very well in behavior and evolution.  And so my colleague, ______, wanted me to come over and give them an overview of the kind of work that was done on behavioral evolution, because they don't do a lot of this behavioral work.  On the other hand, they have absolutely amazing populations, especially of large mammals, of everything you could imagine, from wild asses to wild horses to camels to goats and sheep, that would make wonderful study systems for people who are interested in behavioral evolution.

But again, they just haven't been trained in that.  So I went over and gave a course to about I'd say 40 to 50 graduate students, even some young faculty members.  And I gave them the basics that you might cover in a semester class, but just in course form over two days.

It was really, really a lot of fun and they enjoyed it.  And my colleague told me people were already talking about the kind of things they could potentially look at behaviorally in some of the systems that they haven't studied behaviorally.  So I think it was a success and it was otherworldly in terms of just being some place that far away and that different than what I'm used to.  We went out a little bit outside the city of Ulan Bator and to the mountains.  And I saw some of the Mongolian countryside, which was incredible.  And so overall, I think it was both a productive and a fun experience.

Mirsky:      And what does it actually mean when you say behavioral evolution?

Dugatkin:  Yeah, I mean they basically go out and—let's say they're studying a wild horse population or a wild ass population, or something like that.  They're going out and they're taking information about the basic ecology of the organisms—where are they?  How many of them are there?  What are they eating?  That sort of thing.

Whereas a behavioral ecologist might be asking questions about the kind of social strategies that organisms live when they live in herds or packs like that.  What kind of aggression is going on?  Are there dominance hierarchies?  Are some individuals more cooperative than others?  Who chooses who as a mate?  How do they determine where they're gonna live?  How does information flow through groups?  That sort of thing.

Mirsky:      Which is exactly what the article in the June issue of Scientific American is about, although we're talking about a really more—even more, I think, detailed kind of analysis of the roles of the individuals in groups and how they interact with each other—in the article.

Dugatkin:  Absolutely, absolutely, yeah.  I mean, the social network approach is exactly that.  It's trying to understand that interface between individuals in the groups, if they're in, and how can we understand both sort of what the individual does and how that cascades through groups to create sort of group-level phenomenon.

Mirsky:      So let's talk a little bit about the article.  We don't have to talk about every detail about the article, 'cause the article's there and people could read it.  But let's talk about, first of all, the idea of analyzing animals other than humans in a network context.  Where does that come from?

Dugatkin:  Well, it's a long time in the making.  So a brief two-minute history of the way that people study behavior in non-humans would go something like this:

Initially people tended to think of animals as little robots that were programmed by their genes.  So if you watch them in nature, you could think of them as sort of automatons that were responding to some code.  And in this case, the code was a genetic code.  So something happens and the code says do X, and that's what you do.

Over time people began to realize that that was a very simplistic view, that it certainly plays a role.  Genetics certainly play an important on how animals behave, but it's much more complicated than that.  And so they began to look at the way that social interactions dictate what happens next when individuals interact with each other. 

         So when you study the world that way, you, for example, look for a pair of individuals and you realize that how they behave towards each other is partly a function of their genes, but it's also partly a function of what happened the last time they interacted with each other, or what happened when one of them interacted with somebody else two days ago.  Those things really matter when you're trying to understand why animals behave the way that they do.

Over time people began realize that even that was an over-simplification of what happens in animal behavior.  And that's when the social network approach became much more prevalent.  And the idea here is that it's—and you can't even think about sort of pair-wise interactions—Individual 1 interacts with Individual 2, and the next time they meet they do something different based on experience.

Now the social network approach says everybody in a group may potential affect everybody else, either directly or indirectly.  So what I do right now might be a function of my interactions with you, with Bill and Mary and Sam.  But it might also be a function of Sam and Mary's interactions with each other that indirectly cascade back to me through interactions that they've had with other individuals who I then interact with.

And so it's a view that suggests that information flow within animal groups is really important and really prevalent and that it potentially can help us explain some of the really complicated things that go on in animal groups.

Mirsky:      And it's sort of self-evident when you consider, let's say, an infectious agent, that if a virus can go through a group and the original individual that had the virus can spread it throughout the entire group, with the last member of the group having never seen the original carrier of the virus.  I think people have an understanding of that.

But what you're saying is the way A looks at B, then B saying, "Why do you look at me like that?"  And then B treats C a different way, and C wanders over to over D and says, "Well, I'm having a bad day too," and all of a sudden Z is getting the back of somebody's hand because of the way A looked at B.

Dugatkin:  That's right, that's right.  It's those kind of complex, indirect things that may be much more important than we've given animals credit for.  And even your example of the infectious disease, I think that you're right.  That's a very intuitive one, because there's sort of something that's clearly moving between members of a group.  In that case, it's a disease.

         But if you think about the social networks and behavior stuff really broadly, imagine that that something that gets introduced into the group is a new behavioral tradition.  For example, figuring out how to get at a new food source or how to dig up something that nobody had ever figured out who to dig up from the ground and eat.

         So now you have this thing, a new behavioral tradition, that perhaps one individual, maybe just by luck, comes upon and realizes, hey, this is a really good way to get food, or it's a really good way to avoid danger.  If others in the group are paying attention and looking at this, then that piece of information, that behavioral tradition, can spread through a group, just like a virus could.

And so it's all of these things that people who study social networks now are really becoming interested in terms of how does this structure, what we see out in nature in real animal groups?

Mirsky:      We're both big Yankees fans.  Just look at what's going on.  The Yankees are all growing mustaches.  So that's sort of a ridiculous example of how a behavior can spread through a group.

Dugatkin:  Right, right, but at the same time, is it being kind of this cute example with our beloved Yankees?  Even in non-humans there are examples where behaviors get introduced into groups that on the surface sort of seem to be trivial.  Like there's no reason that the Yankees should be hitting any better because they have mustaches.  I mean, they're the same players they were before the mustaches. 

But if it creates, for example, a group solidarity, if it creates a sense of being part of something that's larger than the individual, then we know sort of from psychology that can have real physical consequences.  And we don't have as much information about that in non-humans, but certainly, for example, there is this wonderful set of studies that were done that have been going on for 50 or 60 years now in various chimpanzee populations, mostly in Africa.

And there are these six long-term studies of chimpanzee populations.  And these behavioral traditions are really important in chimps.  That kind of makes sense.  They're our closest relative.  And they use culture more than probably any other animal.  But they do all these weird things that are like equivalent to the Yankee mustaches.

In some populations, individuals do this ritualized, stereotypical—we have no other word for it besides dance, when it starts to rain.  Nobody knows why the chimpanzees in one of these populations do this.  But somebody started doing it at some point.  It may create bonds males or between mating partners.  And you see it in some populations of chimpanzees.  But if you go to other populations of chimpanzees, you don't see it.

And what the social networking approach lets us do is try and figure out how does that happen?  Are there some individuals, for example, that everybody else looks to as kind of the central individuals in the group who set the trend?  That might be, for example, could be the oldest female, could be the biggest, strongest male, could be something else.

But if there are these kind of central individuals through which information spreads out, then we can use all of these sophisticated mathematical techniques that people in the social network world have developed that allow us to track how that information can flow within a group and maybe even between groups.

Mirsky:      Yeah, it's really pretty amazing.  You talk in the article about the observations of this tribe of macaques.  Now, that's at the Yerkes Center in Atlanta, so it's not a wild situation.  But they're living in a group and they're having legitimate interactions, even though it's not in a completely, quote, "natural," unquote, setting.  And there was one work done that you talk about where particular individuals were removed from the group to see what would happen. 

Dugatkin:  Yeah, this is a great study.  So the macaques live in groups and there are all sorts of complex rules associated with macaque groups.  And it turns out that in groups there are a few individuals that are referred to by animal behavior as the policers.  And what they do is they sort of keep order.  They tend to be higher up in the dominance hierarchy and what they do is they break up fights, for example, when the fights occur between other individuals in the group.  So they're kind of an law and order part of the group.

And what the researchers—this team was led by a researcher by the name of Jessica Flack.  What they did at Yerkes was the equivalent of what geneticists do when they do a knockout experiment.  So what they did is they went in and they basically removed the police, the policing individuals.

Okay, so when they did that, they got a number of things that happened.  The first that happened and the one that would be the most intuitive, was you take out the police, there are a lot more fights.  And this is the sort of thing that you might expect.  But all of these other emergent things came up when they took out the police.  

o for example, individuals played with each other less when they took out the police.  Individuals groomed one another less when they took out the police.  We don't know exactly why, but you could certainly envision hypotheses like playing with each other, you have to have a certain level of comfort in order to feel like you want to play.  And maybe when these policing individuals aren't around, that level of comfort goes down and you're not quite ready to do this kind of relaxing thing because the whole situation is not nearly as comforting to you.

The other thing that happened was that one of the things that social network people do as—they measure the extent to which a group is made up of subgroups.  And what happens when you take out the police is in the article we refer to as this kind of Balkanization.  The group itself still exists, but it sort of breaks up into smaller subgroups.  And all of a sudden instead of one homogenous group, you've got these subgroups that generally speaking are not interacting aggressively with each other, but it's certainly not as kind of a pro-social grouplike situation as you had before.

And of that comes from just removing policing individuals, demonstrating how much these complex, indirect effects manifest themselves.  I mean, there's no way we would have thought that individuals might play with each other less if you took out the police or that they might groom each other less.

You and I right now, I certainly came up with possible explanations for why that might be, but I never would've guessed that would have happened beforehand, because it's not something that jumps out at you until you actually do the social network analysis. 
  
Mirsky:      Right, and you also talk in the article—there's some material about birds—and we'll let people read the article.  That's also really interesting.  But the one that's kind of flabbergasting is the one with the dolphins. 

Dugatkin:  Yeah, this is an incredible study, a set of studies, that of course people have long been fascinated by dolphins just in terms of what we think they're capable of cognitively.  But they're really smart animals. 

And there's a team of researchers led by a guy by the name of David Lusseau.  And Lusseau was watching dolphins over on the other side of the world in New Zealand.  And he was doing some basic social network analysis.  He was trying to understand how information flows through these groups.

Well, to make a long story short, he eventually hooks up with a fellow in Brazil, Paolo _____ Lopez, who is working a population of these bottleneck dolphins in Brazil.  Well, these dolphins do the most incredible thing.  What they essentially do is they interact with human fishermen who are out there just trying to get their food.  And the dolphins and the humans together hunt down the fish. 

So what the dolphins do is they provide these very, very coordinated, specific cues to the fishermen to indicate that the fish that the fishermen are looking for are present.  So they might slap their tails in a very, very specific, understood way.  That means that there are fish that you fishermen want to get.  Throw your nets out now.

And of course the dolphins also presumably get more food by having these fish netted.  They get more food themselves, but the fishermen get more food as well.  And so if you go out there, you can see the fishermen are clearly waiting for signals from the dolphins.  The dolphins clearly are trying to provide these signals to fishermen.  Well, above and beyond that in and of itself incredible dynamic between the humans and the dolphins, it turns out that not all dolphins are particularly good at this, or at least not all dolphins like to interact and provide humans with this kind of information.

So if you look at the I guess it's about four to five dozen dolphins that they've been studying in these interactions with the fishermen, and if you do social network analysis, it turns out that this group of 50 to 60 dolphins is really composed of three subgroups. 

In one of these groups the individuals interact with each other all the time.  All of the help the fishermen.  In another group, individuals in that group interact with each other.  The dolphins do all the time.  But they never interact with the fishermen.  And in the third group most of the individuals do not interact with the fishermen, but there's one dolphin in that group that does.

And so you get this tremendous variation.  Some of these subgroups love interacting with humans.  Others don't do it at all.  And others typically don't, but some individuals in there do.  We don't know yet how all of this is determined.  So we don't know—for example, in the group where all the dolphins are nice to the humans—we don't know if they teach each other that, if they just sit back and watch and learn from one another, if this might have some, perhaps, indirect genetic component?  Unlikely, but it's probably a teaching thing.  They teach each other other stuff, but we don't know if they teach each other how to interact with humans.  We don't know when that weird group where almost nobody interacts with humans but one individual does, what's special about that one individual that it's determined that it's going to interact with humans.

And how does that individual interact with the dolphins that typically like humans and the dolphins that typically don't like humans?  All of these dynamics are the kind of thing that social networks are beginning to allow us to look at.

Mirsky:      Right, and if you don't do the social network analysis though, then you don't know what to look for.

Dugatkin:  That's right.  If you don't do that, if you didn't do a social network analysis in this case, you would certainly see the dolphins helping the humans.  That would jump out at you no matter what.  But if you wanted to understand what on earth could possibly create that kind of scenario, then you got to do the social network analysis, because you would never know that there's clique of dolphins that love each other and love humans.  There's another clique that love each other and don't like humans.  And there's a third clique where most of them love each other and one of them loves dolphins and also loves humans. 

And the way that that information about how to love humans or how not to love humans, moves around, is only possible when you study these things from a social network approach.

Mirsky:      Right, and we may have actually caught the group with the one member that interacts with humans.  We may have caught that group at an early stage in its development where this individual that interacts with humans is going to bring that behavior to the rest of the group.  That's at least a possibility.

Dugatkin:  That is a fascinating possibility, absolutely.  The other fascinating possibility might be the flip side of the coin, which would be I think interesting if everybody else kicked that dolphin out.  We don't know what's going to happen?

Mirsky:      [Laughs] Right.

Dugatkin:  I mean, it's possible that that tradition will spread within the group.  It's possible that maybe they'll kick the dolphin out that interacts with humans.  Maybe that dolphin ___ interacts with humans in that clique where they don't usually, maybe they'll just get up and move to the clique where they do usually.  None of those things—we don't know yet.  And all of them are sort of interesting possibilities and hopefully we'll know more about this through time.

Mirsky:      This really doesn't have anything to do with the network analysis.  But when you think about the dolphin/human interactions, the humans didn't teach the dolphins how to work with them.  The dolphins taught the humans how to work with the dolphins.  The dolphins trained the human fishermen.

Dugatkin:  Absolutely.  It's incredible.  And we just don't think like that.  Certainly there are many people who spend their time figuring out how to train animals.  But as you say, here's a case where we didn't even realize it while it was happening, but they were essential teaching us something.  I assume what the fishermen were focused on right away was getting more fish for themselves. 

But if you step back and you look at it, it's exactly what you said.  It's them teaching us something.  And then this intricate co-species co-evolutionary thing where now all of a sudden we're both presumably benefitting from these interactions and now we can look at the way that they influence each other, we influence them, they influence us.  This is where a lot of the interesting stuff in animal behavior is going on these days.

Mirsky:      And I think anybody who has a cat or a dog or any kind of of other pet, actually, may have had the thought one day that, wait a minute, they trained me how to do, when to feed them, or when to play with them.  I didn't train them.  They've been doing behavioral modification on me the whole time.

Dugatkin:  Absolutely, absolutely, yeah, yeah.  I mean, certainly pet owners are legendary for telling you those sorts of tales.  And now we can be looking at these things in larger groups.  We can actually go maybe—it might be interesting to go and see whether or not these pets that are training us, whether or not in their natural groupings these kind of things go on amongst themselves.  And so it might just be a small step for them to teach us, even if it's unintentional.

Mirsky:      Right.  They have this behavior in their toolkit to begin with to interact with their ___ specifics, but they just transfer it over to their interactions with us.

Dugatkin:  Exactly, exactly, yeah.

Mirsky:      So I should say I edited the article.  And trust me, because I'm an unbiased source, it's a fascinating read.  And Lee Dugatkin, thanks a lot.

Dugatkin:  It's my pleasure.  Thank you very much.

Mirsky:      Again, the article is called "The Networked Animal" in the brand new June issue of Scientific American.  And look for Lee's last Scientific American piece, "Jefferson's Moose and the against American Degeneracy" in the February 2011 issue.  You can also hear Lee and I discuss that article on our previous podcast.  Just Google Lee Dugatkin, Scientific American and moose.  It's a wild story.  Check it out.

That's it for this episode.  Get your science news at our Web site, www.scientificamerican.com, where you can read our article about the discovery of the first known warm-blooded fish.  It's a fascinating and the fish, called an opah, O-P-A-H, is also noteworthy, because it looks like an area rug.  It's a big, round thing.  Anyway, follow us on Twitter, where you'll get a tweet whenever new items hits the Web site.  Our Twitter is @sciam.  For Scientific American talk, I am Steve Mirsky.  Thanks for clicking on us.