The origin of species: the Australian connection

Why is Australia a cradle for the evolution of complex social systems in birds?

5 December 2006

Professor Andrew Cockburn
Professor of Evolutionary Biology
School of Botany and Zoology
Australian National University

Professor Cockburn is internationally renowned for his work in behavioural ecology and evolutionary theory. His long-term studies of Australian birds and mammals with idiosyncratic life histories have provided insights into intractable questions in evolution. He has studied the life history of the small marsupial, antechinus, to test theories of sex allocation, litter size and sex-biased dispersal. He has used the extreme level of cuckoldry in fairy wrens to show how genetic variation can persist in the face of strong selection. He has also sought to explain why Australia has proved to be a remarkable cradle for the evolution of social complexity in birds.

Andrew's profile at the Australian National University
www.anu.edu.au/BoZo/staffandstudents/staffprofiles/cockburn.php

The superb fairy wren on Wikipedia
http://en.wikipedia.org/wiki/Superb_fairy wren

The life of birds by Sir David Attenborough
www.pbs.org/lifeofbirds

Introduction

Dr Hugh Tyndale-Biscoe: Good evening, everybody. I would like to welcome you on behalf of the Australian Academy of Science to the third public lecture in a series celebrating Charles Darwin and Carl Linnaeus. My name is Hugh Tyndale-Biscoe and tonight I am representing the President of the Academy, who unfortunately cannot be present.

This series of lectures has been organised by the Academy; partly to counter the proposal of so-called Intelligent Design as an alternative hypothesis for the origin of life on Earth and to answer the criticisms that have been raised against evolution. To biologists, this isn’t a very serious discussion, but for people in our society it has become a serious matter. So earlier this year the Academy decided to hold a series of lectures in which Australian scientists, who study aspects of evolution, could speak about their research — most of which is done on Australian plants and animals.

Tonight’s speaker, Professor Andrew Cockburn, has throughout his career been studying evolutionary ecology, that is, the study of organisms in their environment and how they are adapted to that environment.

He began his career at Monash University where he came under the influence of Tony Lee, who got him interested in a small marsupial, Antechinus stuartii — as it was called then — which has a very remarkable life history. Andrew has made substantial contributions to our understanding of the way that this little animal arranges its life.

However, tonight he is going to speak on complex social systems in birds, which he began studying after he finished the work on Antechinus. He addresses one of the key questions, if you like, in evolutionary theory: how does altruism evolve in a society of organisms when, according to original classical theory of evolution, all animals should be selfish and, in fact, fighting for their own survival, and not helping others? This has been one of the challenging tests of evolutionary theory and Andrew has made substantial contributions to this in the past 20 years.

So, tonight we look forward to hearing Andrew’s story on Why is Australia a cradle for the evolution of complex social systems in birds?

(Click on image for a larger version)

Professor Andrew Cockburn: When I was asked to give this talk I gave it a fairly high-faluting title, but what I am actually going to talk about is why I have spent the past two decades of my life studying the superb fairy wren. (Two decades is an accurate summary; we are having the 20th-birthday party of the fairy wren project this Friday.)

One of the reasons that I have devoted so much time and such a large proportion of my scientific career to the study of fairy wrens, is that they have progressively revealed a greater and greater weirdness and complexity; and understanding that complexity is a substantial evolutionary problem.

What do I mean by social complexity? What is a 'complex' social organisation?

The majority of organisms live their lives in a comparatively simple way. By and large, they live on their own and organise their lives in the pursuit of their own interests. But in some societies, individuals are forced to form close relationships with other individuals in order to get by. And those associations become the basis for what appear to be highly cooperative affiliations.

(Click on image for a larger version)

Let's start with a classic and comparatively well understood example. I photographed these lions on the floor of the Ngorongoro Crater, in Tanzania - one of the most beautiful places in the world.

These two male lions are members of a permanent coalition; a male lion can't live its life on its own. Young lions are forced to leave the pride at an early age, probably just as they enter adolescence, and they then spend a miserable two or three years being beaten up and living in the cracks between societies. The majority of them stumble along to game parks, where they are shot by rich German and American game hunters, but the ones that survive to about four or five years old, can have a crack at taking over a pride themselves.

But no male lion can do this alone: it needs to form a coalition with a partner or partners. A group of two, three or four male lions get together, take over the pride, beat up the previous owner of the pride and try to kill it or drive it off. This is impossible to do on your own but quite possible if you are young and sprightly and there are three or four of you. (It is very like political succession.)

Male lions then pursue a collaborative relationship which is selfish to the unit but not to the individual. The first expression of extreme selfishness is to hunt down every cub in the pride and kill them - quickly and brutally. This stops the females lactating, brings them back into oestrus and means the young male lions can start to have their own young.

The lions in the coalition share, in an exactly egalitarian way, the reproductive favours within the pride. They are quite happy to watch their male partner copulate with a female; there is no antagonism between males. It is a collaborative relationship, entirely beneficial to the participants.

It is easy to understand a collaborative relationship which is good for both the partners but it is much more difficult to understand a relationship where one partner seems to behave sacrificially, to the benefit of some other individual.

(Click on image for a larger version)

My favourite example that I have personally experienced comes from the same trip to the Serengeti. I was visiting the Serengeti lion project, so I was allowed to camp on the floor of the Ngorongoro Crater and hear lions roaring and zebras whinnying all night. We wandered around the floor of the crater - we were allowed to get out of the car, which people are not ordinarily able to do - and sat down for lunch with Rob Heinsohn, the friend and former student I was visiting.

A vervet monkey strolled out of the forest, walked very, very boldly up to us, and placed its hand on my knee. Now, this was quite a disconcerting thing to happen to a biologist. My first reaction, as a biologist aware of the risk of infectious diseases, was uncertainty because I realised a monkey bite is dangerous; not only because monkeys have very sharp teeth but because they are very closely related to us and they can transmit diseases. For example, HIV is known to be transmitted from primates to humans - however, transmission probably occurred after eating infected primates.

My second reaction was absolute wonder: to have a monkey with its hand - and I use the word 'hand' advisedly - stroking your knee is something that just doesn't happen in the wild. It was very, very exciting.

And I guess the third, vague, thing in the back of my mind was, 'I've spent the past 20 years studying animals with bizarre sexual behaviour,' and so the motives of the monkey came into question.

However, wonder took over and I sat there staring at this monkey, staring at the hand on my knee, thinking, 'Isn't this wonderful?' Suddenly, Rob let out a blood-curdling scream. I swung round to find that the monkey's partner had climbed into the back of our car, opened our Esky and pillaged it of our entire supply of chocolate.

The motives of the chocolate-pilferer are straightforward: to get lots of chocolate. It is behaving selfishly. The motives of the knee-fondler are absolutely remarkable, because it is doing two astonishing things. The knee-fondler is taking an enormous risk, but at the same time behaving exploitatively and deceptively to a member of another species. It was engaging my attention completely, so its partner could pillage our car of chocolate. It was taking the risk that I would do something that humans do - like shooting or eating the monkey - while the partner got the chocolate. It was behaving sacrificially.

Explaining the sacrificial behaviour, explaining the basis of that cooperation, is arguably the most problematic issue in evolutionary biology.

It is particularly difficult to understand this behaviour in societies which are bound by these cooperative arrangements: monkeys typically live in groups. When we think of organisms that we could use to study this phenomenon, we tend to think of primates. We have a very self-centred view of the world; we think we are complicated, we interact in large groups and as a consequence our closest relatives are likely to be the next most complicated species.

But the number of insights that have come from the study of primates is not huge. In fact, the number of insights that have come from the study of mammals in general is not that great because they are remarkably difficult to study.

(Click on image for a larger version)

Probably the most complicated interactions I have seen in mammalian societies occur between elephants in Africa. Elephants live much of their lives in comparatively small groups or on their own, but every day they have to go to waterholes to drink. The interactions that occur at waterholes are absolutely extraordinary.

Sometimes two individuals who are not familiar or confident with each other will stand for half to three-quarters of an hour with their trunks in each other's mouth. And they'll stand there doing nothing else except exchanging obscure chemical signals which are completely inaccessible to us. Clearly, we don't have an olfactory system that begins to rival an elephant's trunk.

In addition, they engage in a large amount of vocal communication, but once again that's inaccessible to us. Much of the vocal communication is subsonic and formed by low-frequency pulses that can be transmitted over many kilometres. So if an elephant is startled at one site, individuals can be alarmed as much as three or four kilometres away because they are hearing the subsonic alarm call. The individuals cannot see each other but they are totally aware of what is going on. And there is some evidence that elephants probably have a social network of 150 to 200 individuals who they can identify using these subsonic booms. Because it is something we can't hear, it is very hard to work with and difficult to manipulate.

(Click on image for a larger version)

So to understand the evolution of complicated social organisation I have chosen to study birds.

Initially, birds seem a poor candidate for complexity for two reasons. First, they tend to be portrayed as living in simple societies - permanent associations indeed, but those societies are usually portrayed, as in these mute swans, as interactions between a male and a female in a monogamous relationship. Birds are supposed to be happy, loving couples, male and female, staring blissfully at each other. (I didn't get the classic shot of the swans forming the heart, because it's all a bit bilious.)

The other reason that birds appear to be a poor candidate for understanding real complexity is that, frankly, they have bad press when it comes to their brain power.

(Click on image for a larger version)

There are all kinds of derogatory terms relating to birds that are used to describe a lack of intelligence - like 'galah', 'drongo' and 'bird-brained' - and none of these inspire confidence.

The argument I am going to develop today is that both these criticisms, of the idea that birds live complicated lives, are wrong.

(Click on image for a larger version)

One myth that we can dispense with comparative ease, is that all birds live as happy pairs in monogamous relationships. We know this to be untrue for two reasons. First, perhaps like humans - where we draw monogamous analogies - birds are very often unfaithful to their partners. Birds frequently engage in extra-pair matings, so much so that it is now considered remarkable when faithful birds are reported in the literature.

Second, we now know that a much larger proportion of birds than we originally thought, live and breed in permanent groups; collaborating in the rearing of young. This situation, where more than two individuals raise a single brood of offspring, is referred to as 'cooperative breeding'.

Researchers have known for many years that the epicentre of cooperative breeding in birds, on Earth, is Australia. To illustrate this: if you were a bird-watcher and you took a stroll around the Shine Dome, carefully inspecting all of the birds in the immediate vicinity, you would have a reasonable chance of seeing about 25 species of birds that breed cooperatively. By contrast, if you were to start out in Lisbon and you walked all the way to Vladivostok, you wouldn't achieve the same number. Many Australian birds breed in groups, while only 10 per cent of birds do it world-wide. In the geographic centres where the majority of research on birds is conducted - in North America and in northern Europe - cooperative breeding is comparatively rare.

These complicated groups in birds are fascinating from a variety of perspectives - not just from the perspective of cooperation but from the way that within a cooperative society exploitation and parasitism can develop. These groups are incredibly good study organisms to understand the fundamental issue of complex social organisation and evolutionary biology. I would argue that this complex social organisation is one of the big questions in evolutionary biology.

(Click on image for a larger version)

To begin the investigation into the myth of monogamy, let's go back to emus.

Emus are a little bit unusual amongst birds because the majority of care is provided by the father. The male provides all the incubation of the eggs and 100 per cent of the care of the offspring. All the female does is lay eggs.

This male is lovingly lowering himself onto half a clutch of eggs. Once the female has completed about half of her clutch, the male covers them with his body and then engages in a fast; slowly starving himself until the eggs hatch. The female continues to augment the clutch by laying eggs and the male - without moving very much - scrapes the new eggs under his body and incubates them as well.

So how does the female respond to this display of extreme devotion? The answer is: not very well. As soon as the male has committed himself to incubation, the female engages in a display that has been called 'deep booming'. Deep booming involves inflation of a massive sac that the females have in their throat, which creates a loud resonant booming noise that can be heard in the forests at this time of year. The deep booming - which is a very well known call from emus - serves a single function. It advertises to the rest of the world, 'I've got my male bogged down with the eggs. If you come and copulate with me, you can sire the second half of my clutch.'

The second half of the clutch is typically extra-pair, while the first half of the clutch is with the original partner.

(Click on image for a larger version)

Let's look at other examples of avian nastiness. This is the laughing kookaburra, an iconic local bird. These two chicks are getting to the stage where they are close to fledging from the nest —they can't quite fly yet, but they are very close. They illustrate something about kookaburra chicks.

One of them is much bigger than the other. They are different in size because one chick hatched before the others. The female starts to cover the eggs with her body comparatively early in the egg laying period, so the first laid eggs are more developed and hatch earlier than the late laid eggs. As a consequence, there is a clear size hierarchy in the brood.

Now, these two chicks are the lucky ones. They are the survivors of a brutal process which affects all baby kookaburras and which is illustrated in the next slide.

(Click on image for a larger version)

Here are two kookaburra chicks, about one to three days old, with two unhatched eggs behind them. The kookaburra on the right is doing what baby kookaburras like to do, which is to bash in the brains of their siblings or at least peck their eyes out. So these lovely little fluff balls spend the first half of their life being extremely violent.

They have a vicious billhook which develops just for the first three or four days of nestling life, before it falls off. The sole function of the billhook is to destroy the chick's siblings.

(Click on image for a larger version)

Even when we take them out of the nest, to take blood samples and mark them, rather than being traumatised by the process they continue trying to kill each other. You can see here the extremely sharp hook on the beak.

One consequence, if this is allowed to go on untrammelled, is that the younger chicks die and there are only a couple of survivors. Usually the biggest chick beats up on the second biggest chick - the young chicks tend to get caught in the crossfire and lead very miserable lives.

The only response that the parents have available to them is to sit on the chicks all the time. (Even though the mother has squelched the young down, they try to kill their siblings. Although, they tend to peck the mother instead of the other young.) The only way that the adults can maintain attendance all the time, to prevent the young from killing each other, is to collaborate in groups. So where there are group-living kookaburras - the majority of kookaburras breed in groups - there are sufficiently large numbers of individuals to collaborate to restrict the inevitable brood reduction. Nasty little things, really.

(Click on image for a larger version)

In some species the dependency on group living has developed to the point where it has become obligatory. These are the birds that converted me to ornithology, white-winged choughs, which I have studied with my colleague Rob Heinsohn for more than two decades now. Once again they are a very common local bird - they breed on the university campus and probably in the vicinity of the Dome - and habitually live in groups, usually of eight to 10, but it may be as many as 20 individuals.

Two independent studies - one by Ian Rowley who stimulated a lot of this work and one by Rob Heinsohn - have shown that choughs cannot rear young if they attempt to do so as an unassisted pair. At least four choughs are required to fledge one young and they are only capable of rearing all of their young if there are around 12 to 16 adults. By the time there are 15 or 16 choughs, if they lay five eggs all five chicks will survive. This survival rate is linear across all group sizes, but there are still some choughs heading towards extinction. (I told a sad story once about a single chough which took over a bird table in someone's back garden and forlornly fed every other bird that came to the bird table, it was so desperate to be sociable.)

This raises an obvious question in terms of social organisation. Given that the birds need to be in groups to breed, how do new groups form? There are many things that can make groups go extinct: shrinking group sizes, being run over by cars and all the other terrible things that can happen to birds. The question is: how do you ever get a new group?

(Click on image for a larger version)

Rob has advanced the interesting idea that choughs devote a large proportion of their time to forming coalitions with other individuals - very analogous to modern politics! So if anything threatens the survival of the group - such as the death of the older birds or a terrible drought like the one we are suffering at the moment - and the group fragments or disperses, the individuals know that they will be part of the team when the new group forms. And by being part of the team they improve the probability that they will get to the top, and take over and triumph.

So how does a chough prove that he or she is a good team member? One way is to prove that you are a good parent.

(Click on image for a larger version)

Choughs spend a lot of time struggling to raise young: this is an adult chough which is lowering a tasty morsel into the mouths of three begging babies. However, being a good collaborative member and pretending you can provision the babies comes at a cost. It's all very well to do it, but what if you can't?

Young choughs, being fairly stupid when they are young, are not particularly good at foraging for themselves. Sacrificing resources to feed another individual represents a major risk and means food not going into your own belly. How do young choughs circumvent this problem?

The answer is that they appear to have evolved deceptive behaviour. Young choughs, particularly two-year-old choughs, spend a large amount of time picking up large food rewards, muscling their way up to the nest and shooing away the bird who has been the sentry on the nest. In a very ostentatious display, young choughs lower the food into the mouth of a begging baby - you can see that the babies beg pretty vigorously - and then swallow the food themselves before going through a great display of preening the babies and saying, 'Aren't I a good parent?' We believe that they are behaving deceptively to prove their worth as partners.

We don't have a lot of evidence for this, but every now and then they are seen doing this. Every now and then the sentry, who usually departs when a new bird arrives with food, sticks around and there is some evidence of punishment.

(Click on image for a larger version)

Choughs are not very good at punishment. They engage in an extraordinary display -which you can easily see locally - called the 'white-wing-tail-wag' where they display their white wings and they engorge their eyes with blood. This is supposed to be such a terrifying thing that the young will never do it again!

(Click on image for a larger version)

However, all these interactions pale into insignificance when they are compared with the sheer weirdness of fairy wren society. This is the superb fairy wren. It is the commonest small ground bird in natural woodland in the Canberra region. The greatest population density anybody has ever found anywhere occurs in the National Botanic Gardens and it is here that we have been working on these birds for the past two decades.

This is a male fairy wren, with beautiful blue, black and navy on its bib - extraordinary colouration.

(Click on image for a larger version)

Unlike many of the social birds, where males and females are comparatively difficult to tell apart, female fairy wrens are dull brown. They're wildly dimorphic. So telling males and females apart is quite straightforward.

(Click on image for a larger version)

Telling individuals apart is more difficult. We use a technique called colour-ringing. This technique was actually invented by the great Australian ornithologist, Ian Rowley, specifically for work on superb fairy wrens; a technique that is now used around the world.

This photograph was sent to us by someone in O'Connor: the joys of digital photography. They saw a fairy wren with bands on, took a photograph, sent it to us and we were able to tell them the complete history of the bird, which we had banded in the nest. It was a little female, called Omwi, who had dispersed from the Botanic Gardens into the suburb of O'Connor.

(Click on image for a larger version)

fairy wrens were the first ever bird to be recorded as cooperative breeders. This lithograph, by the British natural historian and entrepreneur John Gould and published in the 1850s, is a record of that discovery. It is a perfect depiction of cooperative breeding and as far as I am aware, it is the first record in the literature.

The lithograph shows a female together with two males: Gould clearly recognised that two males participate in feeding. Both of the male birds have food in their bill, which is something else Gould recognised, and they are about to go to a nest. There is an added twist in terms of deception and complexity in the image; the nestling is actually a cuckoo rather than a fairy wren. Nonetheless, Gould knew that these birds were cooperative breeders - about 70 years before it was discovered anywhere else in the world.

He depicted it perfectly; cooperative breeding is provided by males. He sees there is only ever one female breeding in any particular territory but there can be as many as five males breeding in a territory. There is a strong male bias in the sex ratio in territories because males live their entire lives in the territory in which they are born, and hence have a fairly easy life, while females are required to disperse. Females are not tolerated by their mothers and they are forced to disperse in order to breed. Many of them die while they are doing so, and as a consequence there is a massive surplus of males in the population. So, while some territories can be occupied by a monogamous pair, other territories can have as many as five males. (In this case there are two.)

(Click on image for a larger version)

Male fairy wrens have two impressive characteristics. They all diligently care for the babies produced in their territory - by defending and provisioning the nest - and they are all sexually active. All male fairy wrens are fizzing with testosterone and produce lots of sperm. These little birds produce more sperm per gram of body weight than any other vertebrate - more than a bull or a boar. They are the most sperm-active birds in the world.

So how does the female respond to the fact that she is living with as many as five males, all of whom are fizzing with testosterone and lots of sperm? The answer is again: not particularly well.

(Click on image for a larger version)

Females typically spurn all the males in their territory and mate with a male living on another territory. The result: 93 per cent of all fairy wren broods contain at least one young that is sired by a male from outside the territory, and three-quarters of all young are sired by these extra-group males. So, the 'average' fairy wren is illegitimate and legitimacy is a rare exception - very unusual amongst birds.

We have now studied the behaviour of several hundred females; we have dissected parentage by using genetic methods of paternity determination. And of the many hundred females we have studied, only one female has ever been faithful to her partner. (We'll come back to that female in a little while.)

We used the most sophisticated genetic techniques to prove the infidelity - but we should have suspected it earlier. Even in the earliest descriptions of this species, it was known that males spend a large proportion of their time flying outside their territory and performing a display to their neighbours.

(Click on image for a larger version)

This is a male about to engage in display. You can see that all of the pale blue patches on his crown, head and back are erect - the feathers are standing out from the body. If you look at fairy wrens in the spring and you are a reasonably adept birdwatcher, you will certainly see this display.

(Click on image for a larger version)

They then lean their body to the side so that the female sees the pale blue patches shrouded by the black, velvety feathers of the rest of the body. The female doesn't see any of the pale brown or white feathers. She only sees the blue and black - it's called the 'blue-black display'.

(Click on image for a larger version)

On about 5 to 10 per cent of occasions the males augment this display by the presentation of a yellow petal or a flower. All very cute and you can see that the yellow petal stands out brilliantly against the blue, black and navy of the body.

All these photographs were taken on the same morning, at a shrub immediately adjacent to where a female was building a nest. The bird in the first photograph has coloured bands Orange-Mauve-Black - one of the four bands is a metal band with a number on it - so we would call this bird Omen; O for orange, M for mauve and N for noir, to distinguish it from blue.

The bird in the second photograph is called Gummage - Green-Mauve-Green. And the bird with the yellow petal is called Garage, or Green-Red-Green.

Our record in a 30-minute observation is 13 different males displaying to a single female. We estimated that over a period of about a month, the female was probably displayed to by 60 to 70 different males, all of whom performed this dance for her and a proportion of whom presented her with yellow petals.

(Click on image for a larger version)

This was a profound worry to us, as we slogged on through the first decade of research, because although we saw these displays thousands of times - we must have seen about 10,000 of these displays now - none of them ever actually led to mating. We began to hypothesise that fairy wrens were parthenogenetic and copulation never took place. It was only recently that we were able to resolve this difficulty.

(Click on image for a larger version)

We did so with the advent of technology, in this case radiotelemetry. This is a female fairy wren with a radio transmitter on her back. These transmitters are incredibly small and weigh about 0.4 g. (You can see the antenna running out the back.) So, we popped radio transmitters on the backs of females and followed them. This work is as dull as dishwater but the exaltation of the big discovery is one of the defining scientific moments of my life! We discovered that, without exception, three days before the female lays the first egg of her clutch, the very first thing she does when she wakes up in the morning is fly from the roost that she has lovingly shared with her partner, in a straight line to the territory of the male of her dreams. She copulates with him and returns home while it is still dark.

The reason why we hadn't seen copulations in response to the displays: the displays take place in the female's territory but mating takes place in the male's territory, and inconveniently for field biologists it all happens in the dark - very, very annoying.

All of this suggests that females control the mating arrangements completely and this is a real surprise. The females have these males that stuff their young full of food, while they go out and mate with whomever they choose. Females have complete control over the process of paternity and complete control over the process of choosing their mate.

This is comparatively rare amongst birds. Generally, sexual choice amongst female birds is primarily driven by the need to find a partner who will provision their offspring. You choose a male on the basis of his real estate, or on his ability to find food and stuff food in the mouths of your babies - or perhaps into your own mouth.

(Click on image for a larger version)

However, there are a couple of exceptions. This is a photograph of the most beautiful bird in the world, the King Bird of Paradise, which I had the good fortune to see on a trip to New Guinea just a month ago. It is a classic example of a bird which has escaped the need for male care.

Females that look after the young entirely on their own are typically found where food is cheap, low quality and very abundant. So in the case of the bird of paradise, where females provide all the care, the young are fed squishy fruit which is hugely abundant and not very nutritious, and hence having a male to provide more fruit is not necessarily going to enable you to have more babies.

As a consequence, females appear to select males based on the quality of their DNA rather than their quality as providers. They choose features which they think indicate that one male is better than another male. Female King Birds of Paradise have a particular fondness for brilliant orange - and this is the most brilliant colour I have ever seen in nature - bright blue legs and weird tails, which are one of the most remarkable structures in the avian world. You can see the extraordinary, almost astronomical-like green spirals at the end of the long bars.

And this is true generally. Frugivorous - fruit eating - birds in tropical forests often abandon male care and emphasise male beauty instead. Bowerbirds, birds of paradise, cotingas and mannikins all represent repeated evolution of this independence. By contrast, it is unusual for insectivores to abandon male care. The majority of birds eat insects and it would always be useful to have an extra 'pair of hands' to catch insects. If you had a male to help you catch insects, you would be able to rear more young.

(Click on image for a larger version)

As always, there are a few exceptions. A local member of fauna which illustrates this point is the lyrebird. In lyrebirds, the female provides all the care; the male just swans around pretending it can imitate chainsaws, throwing a beautiful tail and generally being gorgeous.

How do lyrebirds get round this dilemma? The answer appears to be that they grow extraordinarily slowly. You can see here that lyrebirds are constantly worried about their weight! Lyrebirds grow at about 40 per cent of the rate that you would expect for a bird of their size and as a consequence the demands placed on the female are comparatively small. The female is able to maintain the rate of delivery of food to the chick and as a consequence she can afford to emphasise something else in her choice of mate: the underlying quality of the male.

So, how can a fairy wren, which is a tiny little bird feeding on tiny little insects, get away without male care? The answer is: they can't but they have developed another trick, which provides the solution to that problem.

(Click on image for a larger version)

The basis for our understanding of this trick comes from an experiment we did by accident. We had known for some time that if dominant males had helpers in the territory - subordinate males - they tended to reduce their own provisioning to the brood. What happens if the helper is removed? Would the dominant male suddenly become much more diligent and much more likely to feed the young?

This is a reasonably straightforward experiment to do: you sneak out, catch the helper, stick him in a cage, fill him full of male worms and record the nest. Our prediction was that the dominant male would take up the slack and become a good provisioner.

The result: not a hint of it. The dominant male continued to provision at exactly the same rate as he had done before. But the big finding in the experiment was when we returned the helper to the territory the following day.

Normally, there isn't that much violence in fairy wren society, but when the dominant male saw the returned helper, which it presumed had been slacking off for the last 24 hours, it went berserk. It spent the next 36 hours devoting most of its time trying to pulverise the helper. We actually cancelled this experiment because we were concerned about the welfare of the helpers. But we did repeat it in other stages of the reproductive cycle and found that males only object to the absence of the helper when the helper is supposed to be providing care.

We think that one of the things females have up their sleeve is that dominant males force the helpers to help, so they can reduce their load and spend their time chatting up the neighbouring female. As a consequence, the females can rely on the fact that they have an alternative source of 'child care' and cuckold the dominant.

This is supported by the observation that the best predictor of whether the home male gets any paternity is whether he has helpers or not. If the dominant male has helpers, the female tends to cuckold him completely. All the babies are fathered by the males from outside the territory. If the dominant male doesn't have helpers - if they are living as an unassisted pair - the female usually allows him to sire one or two eggs in the three- or four-egg clutch.

So we think females have to provide concessions under certain circumstances, but if they have helpers everything is wonderful and they can get away with complete cuckoldry.

(Click on image for a larger version)

We believe that, as in the case of birds of paradise, females are choosing males based on something which indicates their intrinsic quality. Female fairy wrens are trying to improve the fitness of their offspring with high-quality DNA from the chosen males. Females all like the same sort of thing and I'm not talking about some subtle inbreeding avoidance or other reason for female choice. Females all think that exactly the same male is the one that they want to mate with.

Four per cent of males get 50 per cent of all of the fertilisations in fairy wren society and within that four per cent of winners some males do extraordinarily well. Our most famous male, Green-Blue-Black, or GBN - we call him the Mighty Gibbon - has in a very short period of time left 530 descendants amongst the 2000 birds whose parentage we have resolved by DNA testing. So 25 per cent of all the young can trace their ancestry back to the Mighty Gibbon. And remarkably, the Mighty Gibbon, who was loved by all females, was the only male in the study that enjoyed complete fidelity from his partner. So something about the Mighty Gibbon was wonderful. He died a few years ago and I expect we won't see his like again.

(Click on image for a larger version)

We think we know how females distinguish between males. Male fairy wrens are typically only blue during the breeding season. At the end of the breeding season they undergo a moult where they retreat into brown plumage and look very much like the females. So if you look at fairy wrens at the wrong time of year you get the impression that because there is one blue male and about half a dozen brown ones, there must be more females than males. In fact, at one stage they were called 'Mormon birds'. But that is all rubbish! It is just that the males have retreated into a brown plumage.

The time taken to return to the blue plumage used in sexual advertisement is massively variable, and that variation is strongly correlated with age and, we think, quality.

(Click on image for a larger version)

This graph represents the cumulative probability that a bird of a particular age class will have achieved full blue plumage. The horizontal axis is the Julian day at which moult completion occurs. (In Julian days, day one is January 1 and day 365 is December 31.)

fairy wrens start to breed in mid-September, so the majority of one-year-old birds turn blue after the breeding season has started. They are very slow off the mark.

Two-year-old birds enjoy a slight advantage over that - moulting a little bit earlier, but some of them are still fairly late.

Three-year-old birds: again there is a quantum shift towards earlier and earlier moult. On the graph there is one three-year-old bird way over to the left. It decided to skip the whole brown thing and moult directly from one blue plumage to another and stayed blue the entire winter. Only one three-year-old bird did it and unfortunately it died; not a particularly successful experiment.

As the birds get older, they are very long-lived, the probability of achieving the fairy wren Nirvana of moulting from one blue plumage to another, becomes higher and higher. Around 10 per cent of males aged five and over go from one blue plumage to another and stay blue all winter. In the older age classes, at almost any stage during the winter, there will be some birds coming back into blue.

It is the ability to moult blue-blue or moult early, that predicts whether males are attractive to females.

(Click on image for a larger version)

This is illustrated on this next slide, which graphs the probability that a given male will gain at least one fertilisation in the course of the season. One is not a great return for several months of effort, but this graph shows getting at least one.

The horizontal axis is a moult date index to cover differences between years. Zero means the bird moults on exactly the average date for the population; you can see the big sample size there, 222. One means that it moults one standard deviation later than the average for the population, and going left you see one, two, three, four standard deviations earlier than the average population.

You will see this absolutely gorgeous curve. For a behavioural ecologist, seeing data like this is deeply rewarding: this is the best fit ever to a logistic progression. For male fairy wrens, if you moult late you have a very low probability of fertilisation; if you moult early, you are more or less assured of reproductive success. (Remember zero is total failure for the year.)

The trouble with this curve is that although it fits the assumptions of the statistical model perfectly, it is not necessarily what we would expect biologically. There is an alternative suggestion that none of these males would get any paternity at all and you would have to be out at the top left of the curve before you even started to score. We could imagine an alternative cause with a probability of zero for a certain distance, then going up to 100 per cent. So why don't we see that type of perfect choice?

(Click on image for a larger version)

The answer, we think, is that choice in fairy wrens is quite difficult. Early moulting males turn blue in February but mate choice is not consummated until September. So the female has to somehow remember, 'He's the one that has been displaying to me for seven or eight months.' By October, when everybody is blue and everybody is displaying to her all the time, she is bombarded with information and she has to somehow sort out the one who has been keeping his act together.

Even more impressive, she just sits around and gets all the information she needs about the males as they come and display to her, and then months later she makes a straight-line flight in the dark to find that male. This is a really difficult problem for the female to solve and we might expect that getting the mate choice right is quite hard. We expect - by simple evolutionary principle - that the moment difficulty arises, other individuals will come along to exploit that difficulty. So is there any evidence that they do so?

(Click on image for a larger version)

The answer, we think, is yes. This is a schematic, which you can think of as three fairy wren territories, showing the boundaries of the territories. fairy wrens are highly territorial and we can map territory boundaries to within about a metre.

One territory has three males in it. The studly male in this group tends to sing from a conspicuous perch in the centre of its territory. It belts out songs during the entire period that females are moving round. (This is the classic dawn chorus. fairy wrens are the loudest birds in the dawn chorus in Canberra in spring.)

The studly male generally has two less studly helpers. The two helpers also sing during the dawn chorus and they sing as close physically as they can to the dominant. So they find a perch nearby and belt out songs themselves, even though there is a 100 per cent chance that every single morning they will be beaten up for doing so. The dominant, almost as a ritual, goes over and boxes them round the ears, before heading back to his perch to sing.

What about low-quality dominants? Low-quality dominants assess themselves as unlikely to attract females, so they go as far as they can to the edge of their territory, lean out on tippy-toes as close as they can to the studly male, and sing from there. We even had an extraordinary case of a territory boundary we had mapped pretty accurately, which had an old acacia branch bending out over the boundary. The base of the tree was in the weak male's territory and he used to go right to the end of the branch, into the studly male's territory, and sing on his tippy-toes from the studly end. We think weaker makes are trying to behave parasitically, to pretend that they are the studly male.

So a female arriving in the dark is confronted, not with one male but with potentially six or seven. One of my students, Anna Dalziel, has also shown that these low-quality males sing, as closely as they can, the dialect adopted by the dominant male. They imitate its song and females have to somehow sort this out.

(Click on image for a larger version)

This behaviour is something that the dominant male dislikes and it generates enormous amounts of competition and anger between the males. I have already mentioned that there is constant fighting and punishment between males at this time of morning. Also, the helpers really push their luck. If the dominant male is distracted and leaves his perch, the helper immediately abandons its perch and hops over to the dominant's perch and starts belting out the song, pretending to be the dominant. So a female arriving, believing that it is the dominant's perch, is confronted in the dark by a little wussy male who is belting out the song and she might as well mate with him.

If this were the case, we would predict leakage of paternity to helpers. And this is, in fact, what we see.

(Click on image for a larger version)

In this graph there are two curves. The curve at the top is the same one I showed you earlier for the dominants. The curve at the bottom is for the helpers, but the helpers are assessed here not by their own prowess, which has no predictive capacity at all, but by the prowess of their dominant male. So, once again the dominance of the male predicts the likelihood of extra-pair paternity. It is not as high, but if you are a helper living with an attractive dominant male, you have a greater chance of reproductive success than if you are an average dominant. We believe this is purely parasitic and the attractive dominants are ceding part of their success to the helpers.

(Click on image for a larger version)

Let's put this all together into an evolutionary scenario. Some of this is conjecture, but most of it is supported either by experiment or by good contextual evidence.

We suspect the first phase of the fairy wren society was the evolution of coercion on the part of the dominant male to enhance their ability to obtain extra-pair mating. It probably gave them more time: if they could get the helpers to do the work, they could spend more time strutting their stuff.

Females used the greater discrimination they gained from having lots of male display, to increase the rate of cuckoldry, which is intrinsically unstable. Under these circumstances you would expect the dominant to defect and not be prepared to care for young which he suspected were not his own.

However, the female has responded by trading paternity, to give the male some hope. Females always mate with their partner when they come home from visiting their neighbour and if there are no helpers, they mate with their partner sufficiently well to allow him to sire one or two of the young. But collectively this has given the female considerable power and has led to extreme mating asymmetry, with some males doing extremely well.

However, the females have dug a grave for themselves. By relying on the long-term cue, which may discriminate adequately between males, and by insisting on control, which means they visit the male in his territory, they have made the system error-prone. This has led to the evolution of parasitic strategies, which in turn stabilises the helping behaviour.

So we see quite a complicated counterplay of the interests of various participants in the society, leading to the society we observe.

It still seems reasonably obvious that there are many ways the system could collapse. The males could detect things going awry or they could try to coerce the female not to depart, and the system might collapse back to monogamy. You could also imagine circumstances where the social system drifts off in the direction of becoming like that of the birds of paradise, but it doesn't.

(Click on image for a larger version)

Even more remarkable: although I have been talking largely about the behaviour of one species of fairy wren, there are many species of fairy wrens and many aspects of their biology are different - except the very feature which sows the basis of the evolutionary instability.

(Click on image for a larger version)

fairy wrens are part of a diverse radiation; they live in virtually every habitat in Australia and Papua New Guinea. Here is a red-backed fairy wren, which is found in tropical grassland; a white-winged fairy wren, which occurs in saltbush in the desert; and a lavender-flanked fairy wren, which occurs in the Kimberley.

There are huge differences in many aspects of the way these species organise their lives, but they are all dependent on extra-pair mating - suggesting that something about this system has become sufficiently stable that it survives speciation events and it persists over evolutionary time. Clearly, an important question is how this comes about.

(Click on image for a larger version)

We know from studies of other aspects of sexual conflict between males and females that it can generate bizarre evolutionary outcomes. Standard sexual selection is a tense process between male and female over mating opportunity, and the universal symbol of this behaviour is the peacock. We know that sexual selection can lead to bizarre evolutionary outcomes - the peacock at one end and the King Bird of Paradise at the other.

Increasingly, we see from the study of the cooperative breeders that social organisations can also evolve chaotically in response to sexual conflict. There is no other species, so far as we are aware, that is even remotely like a fairy wren. There have now been detailed molecular studies of over 30 species of cooperative breeder and it is possible to distinguish 25 distinctive and very different mating systems amongst those 30 species. Where there is similarity, the similarity arises because they are each other's closest relative, which once again indicates that these complicated social systems - which are often very distinctive - are stable once they have evolved.

We think the stability of these systems is driven by the ability of some individuals to do very well. There are always going to be some winners in these systems and because those winners are the ones that provide the next generation, reproductive success is assured.

(Click on image for a larger version)

I will end this talk with some thanks. This work has been very generously supported by the Australian Research Council, and, perhaps even more generously, supported by the Australian National Botanic Gardens, which is surely one of the jewels of Canberra. And it has been supported by a great number of research assistants, postdocs and, in particular, my long-term assistant, Helen Osmond.

Thank you.

Discussion

Question: Do budgerigars breed cooperatively?

Andrew Cockburn: No. Cooperative breeding is comparatively rare amongst parrots; it has only been recorded in about a dozen species. The most famous of those species in Australia is the eclectus parrot, which breeds in a polyandrous association where a female is supported by as many as seven or eight males. But budgerigars are, as far as I am aware, completely monogamous.

Question (Hugh Tyndale-Biscoe): Andrew, you didn't mention what happens in the succession. When that very successful male had 530 offspring, did his helpers gain more success than the helpers of a less successful male?

Andrew Cockburn: I probably gave the wrong impression there. He didn't have 530 offspring; he had 530 descendants, which includes his children and his children's children. We got him about halfway through his life - he was already out there strutting his stuff when we started - and he lived right on the edge of our study area. We recorded about 75 offspring for him, which is more than have been recorded for any other bird in many species in the world, despite the fact that some birds have clutch sizes of ten, and harems of many females. But his children and children's children, and those children's children, have gone on with comparative success, and this is what gives rise to the massive triumph.

There is another bird which has about 500 descendants. Because there has been some intermingling of the gene line, some of those are the same. But after that, the third best has about 150.

Question 3: If white-winged choughs have become dependent on group living, do they have to protect themselves against inbreeding?

Andrew Cockburn: Most birds have strong incest taboos. White-winged choughs are unusual in that occasionally incest has been recorded and that is a little surprising. In the majority of these cooperatively breeding species, there is an absolute inhibition. So how that is resolved is quite different.

One of the original explanations that were suggested for how fairy wren society evolved is that because they are living in family groups, females mate outside the group. But in fact the nature of fairy wren society actually makes incest more likely because they tend to be surrounded by clusters of male relatives that they don't know anything about. Generally, there is a strong observation of incest taboo. Occasionally you make mistakes, but they are mistakes in a context where you couldn't possibly have known that it was a relative. So we have had cases of fairy wrens whose sons have had extra-pair offspring and then the females have mated with the extra-pair offspring of their sons.

fairy wrens have it down pat. For a female fairy wren, if the dominant male is her son and gets to the top in the territory, come mating time he ignores her and she ignores him - no sexual interest at all. So incest avoidance is nearly universal.

The exception is a weird one that occurs in the group called the Gallinules: moorhens, pukekos and rails. My suspicion is that it has evolved in that group as a way of facilitating colonisation. These birds are particularly good at moving from one water source to another, and they are particularly good at plucking small Pacific islands out of nowhere. They tend to fly furiously across the Pacific and land on an island where they immediately evolve flightlessness, and in order to be able to do that successfully they often need to be able to mate with their close relatives, which they do.

Question 4: Why should there be more cooperative breeding in Australia compared with Europe?

Andrew Cockburn: I probably didn't make that point sufficiently clear, because I was running out of time. There are probably several reasons for that. One of the foundations for, I think, complicated interactions that can lead to this type of interaction is a set of circumstances where children grow up and become sexually active in proximity to sexually active parents. That raises the conflict over parentage considerably, because children become competitors with their parents. This occurs where birds are comparatively long-lived - the parents need to live long enough to see their offspring become sufficiently long-lived to become breeders themselves. And under those circumstances they start to play a variety of experiments of resolving the conflict that inevitably arises.

The majority of birds of the upper Northern Hemisphere, in the postglacial wastelands of northern Europe and northern North America, comprise migrants or small birds with enormous clutch sizes, which typically only live for a single year, and therefore rarely coexist with their close relatives. So if there is any coexistence between close relatives, it would be between siblings. But again, getting back to the inbreeding avoidance cost, this is partly resolved by the fact that one sex tends to disperse and the other sex tends to stay more or less in the same place, which means that brothers and sisters can't interact and interbreed.

The competition that exists between parents and offspring doesn't arise in short-lived species. Long life facilitates this type of breeding. Australian and South African birds - I am about to go on sabbatical leave to South Africa - tend to be very, very long-lived. In these environments, the frequency of cooperative breeding becomes very high.

Hugh Tyndale-Biscoe: Andrew, thank you very much for a wonderful talk and the insights into vertebrate behaviour. I think we can learn quite a lot from fairy wrens and perhaps use that understanding to make more use of the cooperative aspects of fairy wren society. Thank you very much.