Dolphins mind the shell: About an innovative foraging strategy that spreads beyond the mother-calf bond

Being part of the ‘Dolphin Innovation Project’, Sonja Wild was privileged to study the fascinating behaviour of the Shark Bay dolphins in Western Australia. To Cultured Scene, Sonja provided some insight into her study on ‘shelling’, a novel foraging strategy that is transmitted via social learning. The whole study by Sonja and her colleagues has recently been published in Current Biology.

The paper

Cultured Scene (CS): What is the key finding of your study?

Sonja Wild (SW): Our study shows that ‘shelling’, a foraging innovation in dolphins, spreads through social learning among associates. These results are surprising, as dolphins – and toothed whales in general – tend to follow a do-as-mother-does strategy for learning foraging behaviour. We therefore provide the first quantitative evidence of non-vertical social transmission of foraging behaviour in toothed whales, where we have also controlled for influences of environment and genetics.

CS: You describe the foraging strategy of ‘shelling’ in Shark Bay bottlenose dolphins which qualifies as an example of tool use. How does this strategy work? Was it ever found outside of your study site and do you know of similar tactics in other animals?

SW: During shelling, dolphins chase their prey into empty shells of giant gastropods (marine snails), insert their rostrum into the aperture of the shell, bring it up to the water surface and shake it above the surface until the fish falls into their open mouth.
There are one or two anecdotal reports of shelling behaviour from around Australia, but there are no known populations in which it occurs as frequently as it does in Shark Bay.

CS: Are there specific circumstances (e.g., environmental) that promote or increase the use of shelling in Shark Bay bottlenose dolphins?

SW: To acknowledge the fact that behaviour can spread through a combination of social learning, environmental and genetic factors, we integrated three different networks in a ‘network-based diffusion analysis’ (NBDA): the first network was based on association data modelling social learning, the second depicted dyadic home range overlaps as a proxy for environmental similarity experienced by individuals and the third was a network with genetic relatedness to control for a genetic predisposition for shelling.

Our results show that the association network best predicted the spread of shelling, indicating that shelling spreads socially among closely associated individuals, while influences of environment and genetics appeared less important.

Of course, shelling can only occur where dead shells of those giant gastropods are available, so an influence of environmental factors cannot be fully excluded. Our models show, however, that environmental similarity does not solely predict the spread of shelling. 

CS: Why is this topic important, and how do you feel it relates to social learning and cultural evolution more broadly?

SW: These findings on social transmission of shelling outside of the mother-calf bond have important implications for understanding how dolphins may be able to adapt to changing environments, as social learning strategies can be used as a form of behavioural adaptation and therefore increase chances of survival.

Theory suggests that stable environments should favour vertical learning from mother to offspring, as the previous generation’s knowledge is tested, experience-based and adapted to the current local conditions. In changing environments, however, such behaviour may become outdated and no longer pay off. Therefore, unstable environments should favour horizontal learning, as it allows a rapid spread of novel adaptive behaviour.

A marine heatwave in 2011 has caused mass mortalities of fish and invertebrates in Shark Bay, reducing prey availability for the dolphins drastically. Interestingly, almost half of the shelling observations were made in the two years immediately following the heatwave. While we can only speculate as to whether the drastic environmental changes selected for a non-vertical spread of shelling, it appears plausible that the increased abundance of dead giant gastropod shells would have increased learning opportunities for shelling. 

“An average day on the water in Shark Bay beats any day in the office.”

The process

CS: The dolphins of Shark Bay (Western Australia) are subject to a long-term study, now spanning almost 40 years of research. How did you find your way into this fascinating project?

SW: I have been conducting research on the Shark Bay dolphins since 2013 when I joined the ‘Dolphin Innovation Project’ ( as a MSc student at the University of Zurich (Switzerland) with Michael Krützen. My initial goal had been to work on primates for my MSc, but I got talking to Michael during one of his courses, where he first told me about this ‘shelling’ behaviour and the possibility of investigating its spread as a project, which I then soon agreed to. After my MSc degree, I joined Will Hoppitt at Leeds University (UK) for a PhD to learn more about NBDA and continue the work I had started during my MSc, as there was so much more to learn about the shelling and other fascinating dolphin behaviour.

CS: I can imagine that data collection via boat-based surveys was very challenging, especially if the behaviour under study can only rarely be seen. How did a typical sampling day look like in a nutshell?

SW: Data collection is very much dictated by weather: on non-windy days, the team heads out on the water searching for dolphins. The focus of data collection has never been on shelling alone. As a long-term project, the Dolphin Innovation Project has had varying foci over the years, from foraging specialization over demography to genetics. We record behavioural data during the first five minutes of each encounter, including group composition and predominant behaviour as baseline data, but then also note any unusual foraging behaviour seen, including shelling. Shelling is indeed easy to miss, as it can only occur where dead shells are available and is typically of very short duration. Over the twelve years, we have only seen 42 shelling events performed by 19 individuals.

CS: How much is known about the dolphin’s family structure and how did you (if possible) identify individuals during your surveys?

SW: The Dolphin Innovation Project was established in 2007 – since then, over 5400 dolphins groups have been encountered and more than 1000 different dolphins identified. As in the Shark Bay dolphins both sexes are philopatric, we tend to know the individuals within our study are very well, including who they associate with, how these associations change over time, what specialized foraging behaviour they engage in, and of course we keep track of new calves being born into the population. We are able to tell individuals apart based on the marks and nicks on their dorsal fins, meaning that upon each encounter, we need to photograph the fins of all individuals present.

Photo ID: Taking photos of a dolphin fins for identification on a glassy day in Shark Bay (credit: Nahiid Stephens – Dolphin Innovation Project)

CS: What were the best and worst aspects of data collection – any funny stories? Any specifics about working with dolphins?

SW: As with any fieldwork, there are fantastic days and those that are not so great. I kept saying though that even an average day on the water in Shark Bay beats any day in the office, so I really cannot complain about my field days!

The best aspects of the fieldwork was definitely the incredibly varied behaviour of the dolphins. From one minute to the other, they could go from resting at the surface to bolting away as there were other males trying to steal ‘their’ female, often resulting in vicious fights with tail slaps and body rams. Another time, we witnessed two adult males ‘harassing’ a sea turtle – they repeatedly swam very close circles around the turtle and blew bubbles at her with the poor thing just spinning on the spot completely freaked out – and all of that just for fun.

In terms of challenges, I’d definitely say the weather and other wildlife. Weather, because against all expectations, Australian winters can be quite cold, and combined with wind and waves on the water, it could make days rather unpleasant, particularly when being caught in unexpected storms. And while most of the wildlife encounters were pretty spectacular, Australia has some nasty creatures to offer. No, I’m not thinking about spiders and snakes – by far worse were the ‘midges’ (sand flies) – they tend to crawl under your clothes, and their bites can swell up to tennis ball size under hot water or when lying under a blanket…definitely caused many painful and sleepless nights.


CS: How did you manage the writing process? Was it straight forward, or were there challenges?

SW: I have been working on the shelling since the start of my MSc in 2013 – accordingly, there have been countless versions of the manuscript over the years. It wasn’t until towards the end of my PhD (end of 2018) – after I had finished data collection for good and refined the methods – that I finally managed to find the right angle and produce a version that myself and co-authors liked. However, between the viva voce, finding a new position and the start of my postdoc, it got thrown on a pile of things to do later and I struggled to get back to it. It wasn’t until nearly a year later that I finally managed to finish those last 10% to get it submission-ready. With nearly 7 years in the making, the process was definitely not that straight forward, but it really paid off in the end.

CS: How was the peer review process?

SW: I have been extremely pleased with the review process: both reviewers have read the manuscript in great detail, have made constructive suggestions and helped to improve the manuscript overall. Getting the reviews back on a manuscript can sometimes be a bit of a nerve-wracking experience, particularly as an early career researcher, as sometimes they can be unnecessarily harsh and disillusioning. In this particular case: kudos to the reviewers for their professionalism and critical, but constructive comments.

What’s next?

CS: Will you be following up on this research? What questions interest you next, based on your findings?

SW: While I have changed study subjects from dolphins to songbirds for my current postdoc project, I am still interested in similar questions. Using great tits as a model system, I am investigating different social learning strategies the birds employ, and how they adapt their behaviour based on changes in their social environment. Changing the study system can be challenging, but it is very exciting to be able to conduct behavioural experiments on the wild birds. The experimental equipment we use allows detailed experimental manipulations, which provides fine-scale and high resolution data that can never be achieved by purely observational studies.

What I would do differently?

Have a plan B!

CS: As early career researchers, we’re always learning. Is there anything you’d do differently in future, based on your experiences conducting this study?

SW: We originally wanted to do some seabed mapping to figure out where those giant gastropods naturally occur and include a measure of shell abundance in each individual’s home range into the analysis. We were going to use a new underwater ROV to record videos of the seabed on transects – due to production delays it only got delivered during my very last field season, and of course it broke after a few days…. What I would do differently? Have a plan B! Relying on new, untested equipment (particularly for your last field season), is never a good idea.

CS: Finally – what do you think are some of the big questions / challenges facing the field of cultural evolution and social learning?

SW: While there has been increasing interest in animal social learning and culture in the past two decades, the field is still relatively new and accordingly, there are a lot of gaps to fill. Most existing studies on wild populations are biased towards taxa where we ‘expect’ social learning to occur and where it is comparatively easy to track (such as primates, cetaceans, birds), but to really understand how cultural behaviour emerges and how it spreads in nature, we need to expand our selection of study species. Of course, there are huge limitations to studying some species in the wild, but hopefully with technological and analytical advancements, we will be able to track behaviour of more and more species in the wild.

CS: Thank you, Sonja, for sharing your exciting research and ECR experiences with us!

The publication

Wild, S., Hoppitt, W. J., Allen, S. J., & Krützen, M. (2020). Integrating Genetic, Environmental, and Social Networks to Reveal Transmission Pathways of a Dolphin Foraging Innovation. Current Biology.

Watch the video abstract here:

Dr. Sonja Wild is a postdoctoral researcher at the Max Planck Institute of Animal Behavior, employed by the Centre for the Advanced Study of Collective Behaviour (University of Konstanz – Germany). Her main research interests lie in the mechanisms underlying the spread of behaviour in wild animal populations and how animals respond behaviourally to changes in the (social and physical) environment. For her MSc and PhD, she investigated the spread of two remarkable foraging strategies in dolphins using social network analysis. In her current project, she is studying social learning strategies and how changes in the social environment can trigger switches from established to alternative behaviour using great tits as a model system.

Sonja’s personal website:

Follow Sonja on Twitter: