Researchers have known for a long time that dolphins associate in something like human “mums’ clubs”. How important is this for successful reproduction compared with the importance of having good genes?
It turns out that both of them matter, but most notably we have found that the females who have successful relatives and successful “friends” are much better at producing calves than we would have expected from simply adding the effects of the relatives and the effects of their peer group.
This is the first discovery of interacting genetic and social effects on reproduction in the wild, and it may help us to understand whether our own behaviour and reproduction are driven by the same factors as other species. The methods can be used in any species, including humans, to investigate the interaction of social and genetic effects on not only reproduction but also disease resistance, academic ability and other traits. The work will also help us in other studies where we are trying to untangle the effects of different threatening processes on dolphin reproduction.
Why did we focus on reproduction first? Well, evolutionary biology is really simple and reproduction is a key component: organisms that are good at reproducing themselves are still represented by their descendants, while poor reproducers are no longer here. So you will probably be surprised to hear that studies exploring “What makes a wild organism produce babies well?” are still in their infancy.
From our own experience as humans we have the idea that raising a family might be helped by two things: having good genes or by having helpful friends. A social group in humans might help with useful advice ranging from what to name the baby and what mum is doing wrong (everything, usually) through to advice that is actually useful, like how to deal with childhood diseases. Many other animal species form social groups too. Maybe they aid (or hinder) reproduction too. Which is more important: assistance and information transmitted by genes or by social interactions? Sometimes this is called the “nature” versus “nurture” debate. Any scientist will tell you that when a debate goes on for a long time it is probably because both sides have some truth in them. The debate rages not only for humans but also for various wild species where there have been studies of genetic and social aspects of reproduction.
However, before our study these aspects had never been analysed simultaneously in one wild species to enable us to compare the effects of nature and nurture realistically. Why did we choose dolphins? Many researchers study primates in the hope that these close relatives will give us clues about our own behaviour, especially the way we deal with each other in groups. But others, including my colleagues, reasoned that if certain behaviours were really very general we should see these characteristics in more evolutionarily distant species like dolphins, which also have complex group behaviour.
How do we study genes, social interaction and reproduction in dolphins?
First of all, we need to know which dolphin is spending time with which other dolphins, and how well each of them is reproducing, over their lifetime. This is not an easy task for dolphins, which are hard to observe closely unless there is almost zero wind. Dolphin lifespans are similar to ours, so it takes long-term dedication to get the necessary data – a real challenge in a world where scientific funding mostly lasts for 1–3 years. Luckily, these data have been collected for almost a quarter of a century by a huge international team of behavioural researchers at Shark Bay in Western Australia, including Professor Richard Connor, Dr Rachel Smolker, Professor Janet Mann and many others.
Next, if we want to see whether any characteristic, such as the number of calves a female produces, runs in families, we need to know the genetic relationships between individuals. If something tends to run in families we say it has high “heritability”, and this probably means it is partly genetically determined. Heritabilities range from zero – meaning that there is no evidence of genetic effects on the characteristic – to a maximum of one. Of course, non-genetic traits can run in families too, such as the tendency of human family members to live in a certain suburb, so we must be cautious when interpreting heritabilities.
Can we put together pedigrees in dolphins to check for the heritability of calving success? Unfortunately, we cannot really make pedigrees. We can identify mother–calf pairs but the females mate with many males and the father does not help raise the calf, so tracing detailed pedigrees is difficult.
By 1994 Connor, Smolker and Mann realised that molecular methods were needed to go further with the study, so they searched for a cooperative geneticist. They finally found me, and since then we have been joined by a wonderful series of graduate students at the University of NSW, starting with Dr Michael Krützen and more recently Dr Céline Frère, who now collaborate from their new positions at the universities of Zurich and Queensland. We use the same fingerprinting of microsatellite DNA used in human police work to identify whose bloodspot is at the murder scene or who was the father of a particular child.
From the micro satellite data we can give each pair of dolphins a “relatedness” estimate: a number that shows how genetically similar that pair of individuals is. With hundreds of dolphins studied, this became a monstrous table and, as so often happens, the molecular method was way ahead of the statistics needed to analyse it. It was only in 2008 that a statistical method was devised to make sensible heritability estimates using microsatellite relatedness.
When she was my graduate student Frère had a brainwave that led to our study of “nature and nurture”. She was planning to use the new method to study whether dolphin calving success was associated with microsatellite relatedness, asking the question: “Do highly related pairs of individuals have similar calving success?”. If the answer was yes then it would appear that good genes helped females to produce and raise a calf until weaning at 3 years of age. Frère’s flash of inspiration came when she realised that our colleagues’ data included another measure for each pair of females: an index of their tendency to interact with one another socially. She reasoned that since the table of social index values had to be the same shape and size as the table of genetic relatedness values for the same females, she could put the social index table through the same new statistical analysis. In this way she was able to generate the first ever combined investigation of social and genetic effects on reproduction in a wild species.
What did we find?
We calculated each female’s “calving success” – the proportion of years in which the female produces a calf of age 3 years (the approximate age of weaning). Of course there was a spread of values – some females pumped out a calf every 3–4 years while others did not perform as well, often losing calves or not even becoming pregnant. We found that female calving success depends on both genetic inheritance and social interactions. Genetic effects appeared to account for 16% of the spread in female calving success while social effects accounted for 44%.
This means that, for any group of females, their total being more than twice as strong as the genetic effect. This was the first time that these two influences on reproduction had been measured in a comparable way in one wild species. Of course you might say, “I would have expected that answer, so who cares?”. And at first we said that too. However, at this point, the power of analysing the two effects in the same way became apparent, because we could look for interactions between the social and genetic influences on reproduction.
Indeed we demonstrated crucial interactions between these two factors. The mean calving success of a female’s close relatives (the genetic effect) and the mean calving success of the same female’s preferred associates (the social effect) had a synergistic effect on calving success. In other words, the females who had reproductively successful relatives and successful associates were much better at producing calves than we would have expected from simply adding the effects of the relatives and the effects of the associates.
What could be driving this interaction spread of calving could be predicted partly from knowing who each females’ relatives were and partly from knowing who her social associates were, with the social effect between the social and genetic effects?
To try to tease this apart, we focused on each one of the females in turn. For each of these “focal” females we then looked at her preferred associates – the females with whom she spent the most time. We found that when the associates were strongly related to the focal female, their own calf production did not seem to have much effect on the focal female’s reproduction (although we also knew from our previous analysis that the genes that they share would have been affecting the performance of all of the females – the focal female as well as the associates). On the other hand, when the associates were not strongly related to the focal female, their own calf production did indeed seem to have an effect on the focal female’s reproductive success (Fig. 1).
What is next in dolphins?
We have yet to unravel further details of this complicated interaction between social and genetic factors. Genes can help reproduction in all sorts of ways, from making the female better at finding and digesting food through to genes specifically involved in the reproductive process.
Social interactions could be related to protection from sharks – it is not an accident that the bay is called Shark Bay. We know that the dolphins are attacked by sharks and seem to alter their association patterns in time of higher shark density. Females might also help protect each other against males of their own species, especially when the female is younger. Krützen used DNA fingerprints to show that sometimes there are inbred matings – BJ and his daughter Joy produced little Laughin.
Frère has shown that younger females are particularly susceptible to inbred matings, and these matings reduce the female’s lifetime reproductive output. Why would mating reduce reproductive output? The answer is that inbred calves are slower to wean, and since females do not have another calf until the previous one is weaned, this means that the female’s lifetime output of calves is likely to be reduced if she produces inbred calves. Hence females need to avoid inbred matings, and maybe their female associates help them to fend off unwanted paramours.
But can a female dolphin recognise a related male and therefore resist mating? Maybe. Certainly I have seen females who look as if they are desperate to get away from a bunch of rowdy males, but would the females be able to identify which males are their brothers or cousins? We have tantalising information about possible kin recognition between males. Two or more males cooperate to “consort” females, and then mate with them if they are the lucky member of the male gang. In evolutionary terms it makes no sense for a male to help another male to mate, unless it benefits the helper male in some other way. One possible explanation of this puzzling behaviour is that the helper’s favour could be returned sometime. Alternatively, the two males could be related, so that a helper male’s genes are being passed on by his successful brother or cousin.
Krützen discovered that some (but not all) types of male alliance are indeed based on relatedness. However, the relatedness is not close – on average it is about the level expected for first cousins, so males would have to be recognising quite distant relatives when they are forming this type of alliance. So for both the females and males we are interested in how they might identify various types of relatives – up to first cousin – in order for females to avoid mating with such relatives or for males to help their male relatives to mate.
Dolphin calves do have opportunities to learn to recognise their relatives through their mother’s social group. After weaning they may also meet their mother’s next calf – their maternal half-brother or sister. And in cases where related breeding females socialise together, their calves might meet their cousins. However, there are limits to this learning. Weaned male calves spend little time with their mothers, and the next half-brother or half-sister is not produced until after weaning. As far as we can tell, calves would also have little opportunity to learn who their father is or who their paternal relatives might be.
It may be that there is kin recognition based on similarity of whistles or chemical cues, either of which could be partly genetically determined and thus offer clues about relatedness. how can this be used in other species? Already our way of using our statistical analysis will be useful for any species with uncontrolled breeding, including our own. It can also be used to investigate the interaction of social and genetic effects on any characteristic: not just reproduction.
We are hoping that some statistical whiz will come along and write even better methods for analysing data such as ours, so that the interaction of genetic and social factors can be analysed in much more detail than we have been able to. The program was meant to analyse only one thing – the genetic effects – so we were pushing it to its limits mathematically.
*Bill Sherwin is an Associate Professor in the Evolution and Ecology Research Centre of the UNSW School of Biological Earth and Environmental Science.
This article originally appeared in Australasian Science
Copyright Bill Sherwin