Evolution of social behaviour

Into the Storm

slide1

Earlier this year I presented some of my findings at the 2016 Australasian Society for the Study of Animal Behaviour Conference in Katoomba. These findinges were based on data collected over the last 2 years during which my field sites have been impacted by 2 cyclones. Below is an adaptation of the talk from the conference.

The Importance of Sociality

slide2There are lots of examples in nature of animals that form social groups. These species gain advantages and incur disadvantages from their social behaviour. For example an advantage might be better predator detection while foraging (known as the “many eyes” hypothesis) while a disadvantage could include higher rates of disease transmission. Studies suggest that the evolution and maintenance of sociality is likely to be influenced by environmental factors. Changes in the environment, like those caused by extreme weather events, are therefore likely to impact upon the social organisation of social species.

For social species, the balance between the advantages and disadvantages of sociality are vitally important in determining reproductive output, competitive ability, foraging success and survival, factors which can ultimately impact on a species’ ability to recover from a major impact.

The Study System

slide3My research focuses on the coral gobies at Lizard Island, Queensland. Coral gobies are small fish, approximately three to four centimeters in length and they spend their entire adult lives within the branches Acroporid corals (corals of the genus Acropora). They suffer high mortality outside of their corals, and as such rarely move between corals once they have established themselves. I have observed up to 16 species of coral goby at Lizard Island which range in social organisation from strictly pair-forming species (which I will refer to as ‘Asocial’ species) to highly social species which can be found in groups of 12 or more (the largest group I’ve found was over 20 individuals).

slide4During my studies, two cyclones have impacted my sites at Lizard Island which has been quite disruptive to my research, but has also presented a rare opportunity to gain an insight into the rarely studied effects of cyclones on social organisation. There is no doubt (unless you’re a cyclone skeptic) that cyclones cause severe damage to the physical structure of the reef. This destruction obviously has impacts on the abundance, diversity and distribution of reef species following the event. For example, obligate reef-dwelling species (species which depend on the structure of coral reefs for protection and food) tend to decrease in abundance while algal grazers tend to increase in abundance. However we know relatively little about how these events affect social structures of reef inhabitants which is a potential driver of the diversity and abundance patterns we observe. As I previously mentioned, social organisation is important in determining factors such as reproduction, foraging success and survival, all of which are critical for a species’ recovery from a major disturbance.

Methods and Results

slide5

We have been surveying sites around Lizard Island since 2014. During these surveys we search each Acroporid coral on our transects for coral gobies. We identify the gobies to species and count the number of individuals living within each coral head (which constitutes a group). We also identify the coral to species and measure it along three axes to estimate an average diameter. I’ve used average diameter in my research so that my findings are directly comparable to previous work which has used this measurement.

The next few slides show some graphs and conceptual diagrams in which I’ve tried to use consistent symbols which I’ll briefly explain. The yellow fish represent ‘asocial’ species (they’re actually pictures of Gobiodon axillaris, a strictly pair-forming species). The green fish represent the social species (these are pictures of G. erythrospilus which is often founds in groups of 3 or 4). I’ve used a little cyclone symbol with an arrow on the graphs to indicate when a cyclone affected the field sites.

slide6We found that social species decreased in group size following each cyclone while asocial species group size remained the same. This indicates that group size decreases observed in social species were unlikely due to direct mortality from the cyclones (otherwise we would have seen a corresponding drop in average group size in the asocial species as well). A year after the first cyclone, the social species had returned to their pre-cyclone group sizes (keep this point in mind as I’ll return to this in a minute). However, a year after the second cyclone the social species had not returned to pre-cyclone group sizes. This may indicate that multiple impacts have longer lasting effects on social organisation.

slide7

Unsurprisingly, we found that coral size had decreased significantly throughout the study. This was the case for for both social and asocial species. The last set of bars on this graph shows the corals that were uninhabited. I’ve included this to illustrate that corals that were uninhabited at the beginning of the study (darkest bar) were of a similar size to the corals that the gobies were inhabiting at the end of the study. This means that at the end of the study, gobies were cramming into small corals that they previously wouldn’t have inhabited.

Let’s return now to that point I made about the social species returning to pre-cyclone group sizes a year after the first cyclone. From the coral size graphs we can see that these larger groups were cramming into smaller corals than before the cyclone.

Why?

I might pause here for a second to explain the underlying mechanism of the hypotheses of social evolution which I have looked at in this study, the ‘ecological constraints’ and ‘benefits of philopatry’ hypotheses. For both of these hypotheses we need to consider the proposition that social groups arise because subordinate individuals make the decision to delay their dispersal (often at a considerable cost to their own reproductive opportunities). The question of why some individuals will delay or forgo their own reproductive opportunities in order to remain within a group is one of the fundamental questions of evolutionary ecology. There are of course other ideas about why social groups arise, but this idea of delayed dispersal is what I will focus on for this study. It is also important to note that these hypotheses are not mutually exclusive and often act together.  So why separate them out? Well, because each hypothesis contains its own set of testable parameters. These parameters can be used to create a statistical model which we test against the real data and we can determine which hypotheses best describe the observed social structure.

Ecological Constraints

slide8This hypothesis looks at ecological factors which might constrain dispersal from a territory such as a lack of available habitat or high predation pressure. In relation to my work, one of the reasons that the social gobies might have re-formed their large social groups in smaller corals could be that they were constrained by a lack of available habitat. i.e. Gobies displaced by the cyclone might have had no choice, but to move into a coral which already had a small group of fish living in it. In this case, we would expect to see that most of the corals would be inhabited because vacant corals would be quickly taken up by gobies dispersing from crowded corals.

Benefits of Philopatry

slide9This hypothesis looks at the idea that animals gain some benefit of remaining on a site that outweighs the benefits of dispersing. For example, the site might be of a high quality which improves the animal’s fitness to survive and reproduce. Dispersing from this site risks, losing this benefit, unless it can find a site which confers the same or better benefits. In my project, it is likely that there was a lot of variation in coral quality following the cyclone. While fish might have quickly moved into whatever shelter they could find, they might have realised later on that their coral was not very good (indicated by the green, algae covered coral in the diagram), and decided that it was more beneficial to vacate their low quality coral and move into a high quality coral (white coral in the diagram) with an existing group of fish. In this scenario, we would expect to find a lower proportion of inhabited corals than we would under the ecological constraints scenario as fish would have vacated low quality corals in favour of high quality corals.

slide10What we found was that after the cyclone, there was indeed a substantial drop in the proportion of inhabited corals. While this doesn’t definitively prove that benefits of philopatry are causing the observed social patterns, it does lend some support to the idea. There was also a drop in the proportion of inhabited corals for the asocial species, but it was not as substantial as that observed for social species. This likely due to a methodological ‘artefact’ which I won’t get into, but suffice to say, for social species, there is some support for benefits of philopatry playing a role in the observed social pattern following the first cyclone. Stay tuned for a more in-depth analysis of this data.

slide11So, in summary, the major findings of this study were that after a cyclone, social species reduced in group size but asocial species did not. A year later social species had returned to their pre-cyclone group sizes, but in smaller corals. There is some evidence that benefits of philopatry are contributing to this pattern. The fact that asocial species did not alter their social organisation could indicate that the asocial strategy is either more robust to such an impact or that it is less flexible. Unfortunately, my surveys were not designed to examine patterns in abundance and I can’t really say whether either strategy is better or worse for recovery following a cyclone. This would be an interesting avenue for further research. Following a second cyclone, social species again decreased in group size, but did not return to pre-cyclone levels another year down the track. This might be because multiple impacts have longer lasting effects on social structure or because corals had reduced to such a small size that they were not capable of supporting larger groups.

Social organisation in social species is influential  in determining survival. The effects of cyclones on social structures has received little attention thus far in the scientific literature. While my research raises many questions, I hope that it can provide a foundation to build upon and move toward  a better understanding of how severe weather events might impact upon social organisation.

I would like to thank my supervisors and field assistants who have contributed to this work. Also a shout out to the Hermon Slade Foundation for funding this research and  the Lizard Island Research Station for accommodating us.

A picture tells a thousand words Jan – Feb Lizard Island trip

I was joined by my sister, Kaz (author of the Madagascan Adventure series) on my latest field trip to Lizard Island. We repeated the surveys that Kylie, Grant and I had conducted last time and ran an experiment to investigate whether a subordinate fish would decide to move out when exposed to an adjacent coral of varying size and with different numbers of fish already residing in it.  I hope you enjoy this visual expose of our time there.

Life at Lizard

The Experiment

For a brief description of the experiment we ran see here.

Day Reef trip

We were honoured to be invited by Anne, Lyle and Alex for a trip out to Day Reef on our day off. It was meant to be a dry day for us, but we were willing to make the sacrifice!

A small win for the PhD!

I had a small win this afternoon!

I’m back on Lizard Island at the moment and over the last few days I’ve been setting up an experiment.

??????????????????????????

This experiment is to try to determine what factors might influence a subordinate individual’s decision to either stay within a group or to move out. I am testing habitat size and habitat saturation. To test these factors I have created groups of three fish (two adults and a subordinate) in medium sized corals. These groups of three are then presented with either a small, medium or large coral containing either 0, 1, 2 or 3 other fish.

Wow! that’s confusing when I write it all out. Here’s a diagram of my experimental design.

Experimental design

Over the last 48 hours Kaz and I have had the first trial running. In trial 1 the group of three fish were presented with an empty coral which was smaller in size. Last night the subordinate actually moved into this smaller coral. I can’t really draw any conclusions from this one trial, but if we keep seeing this happen, it could indicate that the degree of habitat saturation is more critical than the size of the habitat in determining whether a subordinate will stay or leave a group. In terms of group formation, it could indicate that groups are more likely to form when the habitat is highly saturated (i.e. when there are not many vacant corals). That’s exciting for me and my thesis 🙂

More info on my research

My research

Initial project report

Lizard Log series starting here

Lizard pics

Fish tattooing

Fish tattooing

Capturing the fishies

Capturing the fishies

Goby hunting.

Goby hunting.

Kaz shopping for corals

Kaz shopping for corals

Kaz measuring

Kaz measuring

It's not paradise every day...

It’s not paradise every day…

Initial Project Report

Slide1

Last week I presented at the University of Wollongong Post-Graduate Conference. I have adapted my presentation below as it was a good overview of my project to date. By way of some background, each year the biology post-graduate students are set a challenge to incorporate an object or personality into their presentations. This year it was Leonardo Dicaprio, so keep an eye out for some celebrity cameos.

So without further ado, let’s start this talk by having a quick think about why animals form groups. We might imagine that in a perfect world, the ideal way to ensure that you maximise your genetic contribution would be to breed as soon as possible and as many times as possible. This would involve leaving the natal territory as soon as possible to pursue individual breeding opportunities.

Slide2

So why then do we see so many examples in nature of reproductively mature animals which routinely delay or (in extreme cases) completely forgo their own reproductive opportunities in order to join and remain within a group as a subordinate non-breeding member?

The answer is that we don’t really know. We have a few ideas and theories about the costs and benefits of group living, but a general explanation has remained elusive. A major obstacle standing in the way of achieving a general explanation is that there are a lack of studies on marine organisms, which is what my study will focus on.

One of the most promising frameworks with which to study this behaviour is the cooperative breeding framework. This framework contains several hypotheses. I’ll go through just a few which I would like to test and which I will refer to throughout the presentation.

1) Ecological constraints – EC looks at a situation where ecological factors force animals into groups. eg. high predator abundance might cause animals to group in order to obtain a protective benefit through the dilution effect or by making use of discrete habitat patches which provide defence.

2) Life-history – LH hypothesis and EC are closely linked. LH hypothesis looks at inherent life-history traits of animals which might lead to a situation where group formation if more beneficial. For example, animals with long life-spans might cause a breeding habitat to become ‘saturated’. i.e. no vacant breeding ground for new recruits to make use of. In this case it might be more beneficial for the next generation of reproductively mature individuals to join a group and wait for the breeding habitat to become available (avoids conflict).

Most of these studies on birds, insects and mammals to date have either been broad phylogenetic comparisons or fine scale experimental manipulations. Both methods have merit and have in fact been responsible for the huge advances in this field, but there are few studies combining the two approaches. It is important to combine these methods as the broad scale studies can draw correlative conclusions across multiple taxa, but they don’t offer causative explanations. Which is where the experimentation becomes important. However, fine scale experimentation can only focus on a few taxa so the results are often difficult to apply generally across multiple taxa.

So how am I going to approach this question?

I will use a genus of coral reef fishes which show considerable variation in social organisation as a model. Gobiodon species are found in high abundance on tropical reefs. There are over 20 known species.

Slide3

I will start by conducting a broad phylogenetic comparison of the Lizard island population of Gobiodon. This will involve making a genetic phylogeny of the species at Lizard Island (the above picture shows the phylogeny of the Red Sea population). Phylogenies show how closely related species are to each other. Species radiating from a common node are more closely related to each other, than to species originating from other nodes. The nodes represent some common ancestor. Looking at where sociality occurs on this phylogeny is important as we can see whether the behaviour arose at a single evolutionary point or whether it has evolved multiple times. Looking at the tree above, social behaviour does appear to have evolved multiple times.

Using this phylogeny as a base, I can map ecological and life history traits of each species. This will show which traits the social species share and which the asocial species have in common with each other. I will use this information to identify traits to manipulate experimentally to try to induce social behaviour in an asocial species or vice-versa.

I have chosen these fishes because they show great variation in social structure. for example G. histrio is stubbornly pair forming like Romeo and Juliet. While G. rivulatus forms large social groups more akin to the Great Gatsby, although I suspect that there might be some reproductive shares going to subordinate individuals in the Great Gatsby…

Slide4

But back to Gobiodon; they are also highly site attached, which makes observing and cataloguing their social systems, ecological traits and life history traits far simpler and experimenting logistically easier with them. Once they have chosen a coral to settle in, that is where they stay. Even when the water level drops below their corals during extreme low tides, they will hunker down and remain within the coral, exposed to air for a couple of hours. They have a high hypoxia tolerance and air breathing ability which enables them to do this.

Slide5.1

I have chosen Lizard Island in far north queensland as my study site because:

Slide5.2

Click to enlarge

a) it has an exclusive resort where celebs like leo can be found. Unfortunately, they don’t let the researchers stay here. They tuck us around the corner in this photo.

 

 

 

Slide5.3

Click to enlarge

b) most of the species of gobiodon are known to occur here (and possibly a new species). It’s also worth noting the size of these fishes in the picture below. The fish in the second and fourth pictures on the top row are actually sitting on my gloved hand.

 

 

Slide5.4

Click to enlarge

c) There is a well established research station here run by the Australian Museum which makes field work and experiments much easier.

 

 

We started this work by finding Gobiodon colonies around lizard island and capturing, counting and identifying the species. To capture the fish we use a mixture of clove oil and ethanol which anesthetises the fish and then we ‘waft’ them out of the coral. Once we capture the fish we hold them in plastic bags until the end of the dive and then take them to a boat to be processed.

Slide7

To build the genetic phylogeny we need to obtain genetic material. I’ve been taking fin clips from the fish for this. We just snip off about 1/8th of the caudal (tail) fin area while the fish are anesthetised. It’s not uncommon to see these fish with much larger chunks taken out of their fins, usually from conspecific disputes. The fins do grow back quickly so we’re not doing any permanent damage to the fish.

While we have the fish on board and anesthetised, we measure their SL and TL and we give them little tattoos. These are a flourescent elastomer tag inserted under the skin so that we can identify the fish again when we come back in order to determine some life-history traits like dominant turn over rates or growth rates or mortality.

To collect the information about the ecological traits to map onto my phylogeny, I have been seeking out Gobiodon colonies and taking Coral measurements.

Slide8

To measure habitat saturation we have been using x-transects centred on a colony of Gobiodon. We move along each axis of the transect and catalogue all of the corals known to be inhabited by Gobiodon species. We record whether they were inhabited or not, what they were inhabited by, how many individuals are in each coral, the size and species of each coral.

This gives us an indication of how many and what types of colonies are surrounding the focal colony and how much available habitat there is in the immediate area.

I need to complete the phylogeny now to map these LH and ecological traits and see if there is any correlation between sociality and these traits.

Slide9.1

The downside to working in beautiful tropical locations is that they are prone to cyclone activity.

Cyclone Ita came right over Lizard Island in April this year. The photos below are taken from the same sites (left to right) before and after the cyclone. In February, my assistants, and I had tagged about 600 individual fish with a plan to come back in 6 months to re-capture and re-measure these individuals and determine growth rates and dominant turn over rates. I returned in August and found 8 of the original 600 tagged fish.

Slide9.2

But moving on, I am still interested in the evolution of social behaviour in these fish, but I will focus more on the evolutionary advantages of sociality or asociality in re-colonising a reef after a disturbance. And I’m hoping that I’ll be able to see that recovery in the data coming out of the x-transects that we’re using to measure habitat saturation.

Anecdotally, there appeared to be more uninhabited corals than there were in February, though I can’t verify this statistically because I used different methods (we were looking at a different question in February). There also appeared to be more juveniles present in August than in February.

Slide10

Moving on to some preliminary results, these tables show the results from a statistical method called a Generalized Linear Model or GLM for short. Don’t worry about that or all of the technical looking numbers, all you need to know is that a significant result is indicated in red or a highly significant result in yellow. For most of the species above, there is a significant result for average diameter of the coral. That just means that there was a strong relationship between the size of the coral (the predictor) and the number of individuals living within the coral (the response). i.e. the size of the coral could be used to predict the number of fish living within it for those species with a significant result.

I’ve found that the group size of some of the social Gobiodon species is related to the coral size, but not to the size of the largest individual (alpha), which is interesting as Marian Wong and Pete Buston (who presented at UoW a couple of weeks ago) had found that there was a relationship between both coral size and the size of the alpha with the group size in the anemone fish Amphiprion percula. G. oculolineatus does appear to follow this pattern. What I can take away from this is that the determinant of group size is probably species specific and will therefore be more difficult for me to make general conclusions about.

Looking at some of the before-after cyclone data that was comparable, the corals that did survive the cyclone showed positive growth. However when we looked at the site as a whole, the average size of the corals had decreased. This was to be expected since, as you can imagine, a major disturbance like a cyclone would smash up the larger corals into smaller corals. The smaller corals also have less surface area so are more likely to survive a cyclone.

Slide11

I’ve also found that, as you would expect there was a decline in the coral goby abundance. However, the second graph is more interesting. Some species, like G. erythrospilus, G. rivulatus and G. unicolor appear to be occurring in larger groups post-cyclone. This is possibly an indication that they are in a phase of recruitment.

Slide12

This will require further exploration so I have another trip planned for January. During this trip I will be re-conducting the x-transects in order to examine this trend across multiple sites. What I will be looking for is whether there is a detectable shift in the goby community. There might be a higher proportion of social species present which could indicate that social species have some kind of advantage in recolonising a reef after a disturbance (or vice-versa).

Slide13

I will now need to finish the genetic phylogeny and map on the ecological traits that I have collected so far. I have another round of field work booked in for January. I will be conducting more of the cross transects to see if there has been a detectable community shift since my last visit. I also want to set up a pilot experiment looking at the effects of habitat quality and habitat saturation on a subordinate individual’s decision to move or not.

I would also like to set up and run the life history traits work again. i.e. the capturing and tagging component, as this is a really important part of the cooperative breeding framework which I’d like to explore.

To finish up I’d like to say a final thank you to my assistants for their help in the field. It really is a big commitment for them to come and help me out for weeks at a time. Thank you very much! My work could not happen without your help.

Slide14

If anyone has any questions about my project, please leave a comment below. Thank you!