SCIENCE AT THE SHINE DOME canberra 6 - 8 may 2009
Symposium: Evolution of the universe, the planets, life and thought
Friday, 8 May 2009
Professor Steve Simpson FAA
School of Biological Sciences, University of Sydney
Steve Simpson is a biologist who has pioneered developments in nutritional physiology, ecology, and behaviour. His aims to understand swarming in locusts and to devise a new framework for studying nutrition have provided an understanding of locust swarming that links neurophysiological events in individual insects to mass migration, and led to new insights into the dietary causes of the human obesity epidemic. Along the way, Steve and his collaborators have made contributions to aquaculture, conservation biology, ecology, evolutionary biology and gerontology.
Steve completed an honours degree at the University of Queensland, before undertaking a PhD at the University of London on a University of Queensland Travelling Scholarship. He then spent much of his career (22 years) at Oxford University, where he became professor in the department of zoology and curator of the University Museum of Natural History. In 2005 he returned to Australia as an ARC Federation Fellow in the School of Biological Sciences at the University of Sydney. Stephen has been visiting professor at Oxford, a fellow of the Institute for Advanced Study (Wissenschaftskolleg) in Berlin, distinguished visiting fellow at the University of Arizona, and guest professor at the University of Basel. In 2007 he was elected a Fellow of the Australian Academy of Science, and in 2008 he was awarded the Eureka Prize for Scientific Research.
The evolution of behaviour: From neurones to societies
Steve Simpson: Thank you very much. Today I am going to talk about the evolution of behaviour, and Olivia [Judson] has set me up very nicely. Behaviour is perhaps the prime mediator between the genes of an animal and its environment – it is behaviour which is visible to natural and sexual selection and not genes or their immediate products.
(Click on image for a larger version)
Selection ‘sees’, for example, the speed and manoeuvrability of a cheetah, the intricate behaviour of a male lyrebird courting a female or the detailed, coordinated activities of a social insect group or a meerkat society. But the fact that behaviour is what is seen by, and must evolve through, selection poses a major problem.
(Click on image for a larger version)
Perhaps the central problem in the evolution of behaviour, which is going to be my main theme today, is: how on Earth do they evolve? Behaviours are complex, interdependent traits; how can they evolve? It is made even worse – it was considered virtually intractable until not that long ago – when you consider that behaviour is the product of the brain, and nervous systems are the most complex things in the known universe. So how can this happen?
The answer has pretty much come through in the last 10 years and is related to the stupendous complexity of the nervous system. The answer is that, to a very large degree, the brain during its development builds itself. The neural circuitry of brains is so abundantly complicated that the only way they can self-construct is through activity-dependent processes. Even after a brain has been largely wired up during the later life or postnatal life of an animal, there is still the capacity to functionally rewire it, both chemically and through learning. That means that, in relation to the evolution of behaviour, you can have a relatively simple genetic mutation being accommodated and resulting in a fully coordinated behavioural outcome without requiring simultaneous mutations throughout the entire machinery of the behaviour. A correlate is that the same elements – be they molecules, neurones, muscles, sense organs or what have you – can be recruited very simply to serve very different behavioural functions.
(Click on image for a larger version)
Here is one of the famous examples of laboratory freaks. This was an experiment done by Law, which was originally published in the late 1970s. He implanted an eye primordium into an embryonic frog. When that frog finally achieved adulthood, it was entirely trioptic. It had three functional eyes, with a brain wired up – uniquely for a frog – to deal with that threeeye condition. That is a laboratory freak, but it is illustrative of the power of the brain to deal with a new sensory input.
(Click on image for a larger version)
There are also natural examples, and I am going to give you two – they are two of my favourite ones, I think. One is the evolution of the capacity of owls, at least five times independently, to localise sound in both a vertical and a horizontal plane. Your typical bird can only localise sound along a horizontal plane without bobbing its head up and down, and it does that by using time and intensity differences between its two ears. The brain deals with those two sources of information in parallel, resulting in the construction during neural development of a computational map of sound localisation. What has happened in owls is that they have evolved vertically asymmetrically-positioned ear holes, which has provided a new source of information such that now, in the brain of an owl, the vertical component of sound localisation is dealt with by the intensity-difference channel and the horizontal component is dealt with by the time-delay channel. The same fundamental circuitry in the brain as in other birds has received a new input and has dealt with it. If you look within the brain of an owl, you will find fully functional computational maps of sound localisations in two dimensions.
(Click on image for a larger version)
My second favourite is the evolution of trichromatic vision in some old world primates. Most mammals only have a red (long wavelength) and a blue (short wavelength) channel underlying their limited colour vision. But, in the case of some old world primates, a couple of new world primates and also some marsupials, there has been the evolution of a third colour channel, a green channel. In the case of the old world primates, that seems to have resided in a gene duplication event on the red receptor gene. Just by chance, that gave rise to a new receptor that responds best to a green sort of wavelength band. The brain has just dealt with that new source of information and wired itself up to produce fully functional trichromatic vision.
So an individual can have a relatively simple mutational event, accommodate this change through its development and end up with a new behavioural capacity which, if adaptive, can be recruited subsequently into the evolution of that species. How though do we get from individuals to coordinated behaviour in an insect colony or a meerkat society? How does that happen? I am going to address that question by reiterating some of the earlier points that I have just made and taking you to my particular love, which is, if you like, my barnacle. Darwin wrote and spent a great deal of time working on barnacles. This is my barnacle; it lives, in this case, in North Africa.
(Click on image for a larger version)
We have here a swarm of desert locust. This is a very standard grasshopper, except for one extraordinary respect: it is not just one grasshopper; it is actually two.
Click image to Play Movie
It has a genome within which are packed two different grasshoppers.
(Click on image for a larger version)
If you look at the insect on the left, she was reared on her own; the one on the right is a sister who was reared in a crowd. The one on the left is well camouflaged. It would hide away and be repelled by other locusts. The one on your right, however, is brightly coloured, would aggregate with others and would contribute to vast marching aggregations as a juvenile and to winged swarms as an adult. The behavioural transition between those two forms can happen extremely quickly. Were you to crowd the green one, within literally an hour or two it would be behaving in a crowded manner or ‘gregariously’. So is there something very special or uniquely ‘locusty’ that has evolved, which leads to this extraordinary capacity; or is it simply the recruitment of standard grasshopper features that have been elaborated in particular functional contexts?
(Click on image for a larger version)
One question you might ask is: is there some special stimulus provided by locusts in crowds, which is possessed by no other grasshopper and leads to this outcome? Maybe there is some particular set of visual stimuli or pheromones – odours – associated with locusts. Perhaps there is some special way that locusts have evolved to touch each other in crowds or perhaps it is the way they sound.
(Click on image for a larger version)
It transpires that touch is important and, in fact, there is nothing special about it at all. You can evoke the change in behaviour by stroking a locust on a hind leg with a paint brush. It is the only part of its body that leads to that outcome, but the locust hind leg is no different from the hind leg of any other grasshopper. If you look at the sensory hairs on the hind leg that are responding to the touch, they are essentially no different either. So perhaps it is something to do with the neural pathways that are coming from the hind leg that are somehow uniquely locust? Well, no, they are not.
(Click on image for a larger version)
If you start to enter the nervous system and you cut or electrically stimulate nerves, you can track the information from touch-sensitive hairs, which you can stimulate with a paintbrush, all the way back through the nervous system. There is nothing special about them. There is nothing particularly different between a locust and any other grasshopper.
But, if you think about it, this transition, because it is happening so quickly – within an hour or so of being crowded – cannot involve the nervous system rewiring itself physically; it must involve the same circuits reconfiguring themselves functionally, perhaps under the dictates of some sort of modulatory chemical. Perhaps then there is a special modulator that is released by locust nervous systems in response to being tickled on their back legs?
(Click on image for a larger version)
We have discovered recently that, no, there is not a special modulator; it is serotonin – serotonin, which is the ‘happy’ drug found within the nervous systems of virtually every animal in the animal kingdom. So there is nothing special about it. It is released in that particular part of the nervous system and it evokes this fantastic behavioural transition by acting upon neural circuitry, but it is not special necessarily of itself to locusts. It is so unspecial and so similar to ecstasy and anti-depressant drugs and their action on our nervous systems that you can imagine the sorts of headlines that arose out of the research.
(Click on image for a larger version)
Perhaps, then, a special neuron that releases serotonin has evolved in the locust brain? Again, no. If you start to map the nervous system and the specific neurones that are involved, you find that they are the same old plan of serotonergic neurones – this is within part of the nervous system that we are talking about, in the ‘chest’ of a locust – found in any old grasshopper. The one here [pointing] is, we think, the neuron responsible, but it has just changed its level of activity and has not evolved as a specific new neural system. So what we have here is the evolution of a dramatically different set of behaviours through recruitment of standard mechanisms that were already there in other related grasshoppers.
(Click on image for a larger version)
Another really interesting component, which I think has a much broader interest in evolutionary biology more generally, is that locusts can influence the next generation through their own experience without evolving genes directly. So, when the mother lays her eggs deep in the soil – if she has been crowded recently, she has a memory of when that was – she adds a chemical to the eggs; that chemical diffuses into the eggs and turns the developing embryos into the crowded form. That is the equivalent, if you like, of transgenerational cultural transition in a locust. The locust mother is benefiting from her experience. She has a physiological memory which she is transmitting to the next generation.
(Click on image for a larger version)
When we started to look at what that chemical was, by subjecting it to standard analytical chemical techniques, it turns out not to be anything unique or different either. It is just an alkylated L-dopa analogue that is having this effect. Again, it has been elaborated and amplified in this particular context but, of itself, it is not a unique innovation.
(Click on image for a larger version)
But let us start to consider what is happening at the next level up, which is the interactions between individuals, and start to work towards social interactions. The first question we want to ask is: why do they come together in the first place? The answer to that at the mechanistic level came from modelling the fine-scale distribution of resources in the habitat. If you do that, you can find that they come together because resources become locally clumped. So, they get brought together, against their inherent predisposition to avoid each other; and having been brought together, they jostle; serotonin is released and away we go.
The function of why they do this – why they have very different aggregating and dispersing forms – turns out to be related to predation. If you apply percolation theory, you can see that it helps at high density to break down foraging patterns of predators. So predators are unable to find a highly aggregated group; whereas, if they are abundant but evenly distributed, predators can track their way, endlessly, feeding as they go.
Click image to Play Movie
So we have a mechanism and a function, which has brought together animals – locusts on resources – such that they have gregarised and now they are starting to mill around and do nothing in particular. But suddenly, as if of a single mind, they start to move collectively. They march through the habitat rather like a river of locusts. That is a sharp phase transition: you go from not doing it to suddenly everybody aligning and marching together. Does that imply that there is some sort of higher-order organisation within the group or not? The answer is: no, there is not.
(Click on image for a larger version)
By turning to statistical physicists and using self-propelled particles models, which allow you to model the interactions between entities according to local interaction rules, we found that the transition to collective movement happens as a by-product; it is a mathematical phase transition of individuals following a very simple local interaction rule – and that is, when my neighbour moves, I move too. When you do that, you get these majestic and potentially destructive mass movements.
Click image to Play Movie
You can follow the individuals within arenas as they are going around and around and around in circles and you can quantify these interaction rules extremely precisely. You can then start to ask the question: what happens if we adjust the strength of social interactions within these models? If we adjust them just slightly, does it influence the behaviour of the entire group? The answer is to that is: yes.
(Click on image for a larger version)
If you start to fiddle with very simple local interaction rules, you can get the entire mass of millions of locusts to behave extraordinarily differently as a group. They will form long phalanxes or they will form great big single advancing fronts. They will move at different speeds in different directions and so forth. You can then start to track those sorts of interactions, as we are doing with colleagues in [the Australian Centre for] Field Robotics at [the University of] Sydney by using little semiautonomous helicopters.
(Click on image for a larger version)
So we have the evolution of mass migration behaviour through, again, simple local interactions. But why do they do that at all? What is the point of aligning with others? The answer comes down to something really very sinister indeed, which was demonstrated through looking at another mass migrating animal: the Mormon cricket of Utah.
Click image to Play Movie
It transpires that Mormon crickets, like locusts, are on a forced march to find protein. Basically, the reason you align with your neighbour when they move is that, if you do not, you get eaten.
(Click on image for a larger version)
So cannibalism is driving mass migration, and cannibalism is driven by shortage of protein in the habitat.
(Click on image for a larger version)
And, were you to fall over, you too would be cannibalised, I would suggest.
(Click on image for a larger version)
What we have here is a whole series of simple, pre-existing mechanisms within individuals, or interaction rules between them, which can yield state changes in organisation, which can themselves become subject to selection.
If we start to consider what perhaps I think is the final frontier in the study of biology, it is how to put this theme into the bigger context of biological systems. We are getting a very good idea now of how genes play out through development to produce behaviour.
And we are starting to understand how behavioural interactions can lead to integrated behaviour of groups, populations and, ultimately, communities and ecosystems. But this is not a oneway street; it is girded by abundant feedbacks. What we need to do now, I think, is to understand the relationships between all of those components. So, at the present, there are developing – or already well-developed – bodies of theory that focus on interactions at these various different levels, but we need something that pulls the whole thing together.
(Click on image for a larger version)
I would argue that the focus of attention in achieving such a synthesis has to be behaviour because, as I have said, behaviour is subject to selection. I will leave it there. Thank you.
Discussion
Ross Crozier: We have time for a couple of questions.
Question: Thank you for a very exciting story. It is a fact that, once a locust swarm moves on, there are always solitaria left behind – just the odd one; but all the rest will have gone on and you can find them. Would they be the ones that had the legacy from their mother?
Steve Simpson: They are often just the ones that have been left behind because, if you leave them behind, they revert; they revert as a function of how long they were previously in a crowd. That is where we think you are starting to see the translation from modulation of circuitry through to structural changes in neural circuitry. The longer you are in a crowd, the more your nervous system gets hardwired through gene expression, construction of new proteins, and development and strengthening of synapses. So, if you have been in a crowd for a very long time and you are on your own, you will still behave gregariously and try to find the crowd again – try; otherwise, when you get separated, you will revert back to the solitary condition more quickly. Then you become quiet and hide away, which means that you are less likely to find the rest of the crowd, so that the entire population disperses. If the environment is going to support them, for example, in protein, they become well fed; individually, they walk less; they are less likely to cannibalise, which is less likely to cause mass collective movement; and the population will disperse and they will turn back into normal old grasshoppers. That is the solitaria population.
Question: Steve, you gave that horrific tale of the animal which stops and gets eaten. What happens to the one that stops to eat?
Steve Simpson: That is a very good question. You only stop as a function of the value of the resource, because stopping is dangerous. You can show that, if you provide food – we have done this with artificial foods – and place it in front of moving bands, the probability of stopping and staying in the face of potential cannibal attacks is a function of the quality of that resource relative to your current needs. So, if you are very deprived of protein, you will withstand the risk of predation: you will kick and be very responsive to others but, nevertheless, you will stay and eat because it is terribly important to you. If you experimentally denervate the abdomen, so that they cannot feel their back end, they will sit there happily eating while somebody is happily eating the back of them. You can get daisy chains, which is quite exciting.
Question: When you talk about behaviour in, let’s say, animals, would you say that there is the same continuity in evolution, in a Darwinian sense, as there is for other characteristics?
Steve Simpson: I think there is an even better continuity actually at the level of mechanisms. One big takeaway message is that the same neural circuitry can do many different things. You would even argue that there is pressure against changing the way in which a nervous system functions because, if you start messing around with it, it will become either inefficient or ineffective or it will just catastrophically fail. So, if you have a similar set of neural interaction principles and of underlying molecular mechanisms, you can evolve behavioural complexity upon that same ground plan. And you see this. If you look, for example, at the molecular evolution of voltage-gated ion channels, which are the core innovation – I would argue that they are one of the top 10 innovations in the evolution of life – they collectively determine the behaviour of nerve cells. You can follow their lineage a long, long way back. The same families of ion channels have an evolutionary history that precedes much of metazoan let alone behavioural biology. When you come to the principles whereby neural circuitry operates – the interaction rules at synapses, the way in which nerve cells develop and so on – they too have a very ancient and fundamental set of developmental principles. As with the chemical components – Peter Medawar said something to the extent that evolution is nothing more than the changing use of the same sets of molecules – I think the same is true of neural circuitry. Evolution can yield these extraordinary behavioural differences and can add on all sorts of extraneous things on the outside, such as colours and shapes and forms, but ultimately it is underpinned by a unified and gradually changing set of mechanisms.
Ross Crozier: I think we had better leave it there to get to tea. But I might comment that I think you have brought to us a revised form of the red queen hypothesis, in which of course it was originally said that you have to run twice as fast just to stay in the same place; here, you have to run twice as fast just to stay alive.
Steve Simpson: To keep your backside at least, anyway, yes.
