Now Reading
More Fun Than Fun: The Delicate Truce Between Cleaner Fish and Their Clients

More Fun Than Fun: The Delicate Truce Between Cleaner Fish and Their Clients

A giant moray eel getting cleaned by a cleaner wrasse (Labroides dimidiatus). Photo: Silke Baon, CC BY 2.0

  • An elaborate system exists in the ocean in which some fish present themselves to be cleaned at designated ‘cleaning stations’ visited by cleaner fish.
  • For an evolutionary biologist, competition is expected whereas cooperation is a paradox, whether in animals, humans or even plants or microbes.
  • Competition is easier to explain based on the theory of evolution by natural selection. But cooperation, and especially altruism, requires special explanations.
  • And it is the need for special explanations that make cooperation and altruism most interesting, and attracts so many of us to take on this challenge posed by nature.

One of the more endearing phenomena we see among animals is the cooperation between different species, especially when one partner is big and strong and the other is small and weak. A familiar example is an association between two species where one cleans another by removing ectoparasites off the other’s body. Both parties benefit – the cleaner gets something to eat, and the cleaned is rid of bothersome parasites. You may have seen birds riding on the backs of mammals and pecking on their bodies, even inside their ears and dangerously close to their eyes.

A more elaborate system of such cleaning mutualism is seen in the ocean, where some fish present themselves to be cleaned at designated cleaning stations visited by cleaner fish. These associations between cleaners and their clients are not random, opportunistic encounters but highly evolved and predictably repeated mutualistic interactions between specialised species that know how best to get cleaned and specialised cleaner species that know how best to clean.

There is growing evidence that being cleaned is good for the clients. If cleaners are removed from cleaning stations, clients build up high parasite loads. The most common ectoparasites that clients carry on their bodies are isopods, a kind of lice that specialises in parasitising marine animals. There is also good evidence that cleaners do clean – analyses of their gut behaviour and their gut contents confirm this.

As we will see below, this evolved relationship between cleaners and their clients is a fine balance between cooperation and conflict.

The bluestreak cleaner wrasse (Labroides dimidiatus), which cleans several species of clients, is one of the best-studied cleaner fish. Much of what we know today about the ecology and evolution of this mutualistic association comes from the many years of studies by Redouan Bshary, a professor of ecology and ethology at the Université de Neuchâtel in Switzerland, and his students and colleagues.

Redouan Bshary, in his favourite habitat and attire. Photo: Redouan Bshary

Bshary conducts much of his field research at the Ras Mohammed National Park in Egypt, situated at the southern tip of the Sinai Peninsula, with the Gulf of Suez on the west and the Gulf of Aqaba to the east, and rich in coral reefs and associated fauna. He conducts his laboratory experiments in the Lizard Island Research Station in the Great Barrier Reef in Australia.

His field research methodology is quite enviable: it involves scuba diving with an underwater writing slate, pencil and stopwatch and recording the goings-on at the cleaning stations – who cleans whom, how often, how well and whether the clients are well-behaved. His laboratory research methodology is also clever and innovative, but we will come to that later.

His research has revealed that cleaning-fish mutualism functions as a free market, where clients come to get the service they desire and cleaners provide that service. The market analogy runs deep. Both parties generally cooperate, although they both cheat sometimes as well. If the cleaners merely ate up the ectoparasites and any dead skin, that would benefit the clients a great deal and the cleaners only somewhat. The cleaners would instead prefer to eat the protective layer of mucus rather than the ectoparasites, but this is not good for the clients. If the cleaners eat mucus or healthy tissue, the clients can retaliate in response to such cheating. If the client is a predatory fish, it can, of course, eat up the errant cleaner.

Redouan Bshary’s field study site at the Ras Mohammed National Park in Egypt. Photo: Redouan Bshary

Most clients who avail the services of cleaners are not predatory, but they do have the option to punish them by chasing them around or leaving to avail themselves of the services of cleaners at another station. It is, therefore, in the mutual interest of both clients and cleaners to build a relationship of trust.

When the cleaners cheat and bite off some mucus, the clients notice this and give a jolt. Perhaps this tells the cleaners that their illicit behaviour has been detected, and the cleaners then seem to appease the clients by stroking them in a distinctive way, something like a reassuring massage.

Most of the time, there is good cooperation: the cleaners refrain from eating their preferred food (mucus) and make do with their less-preferred food (ectoparasites). The cleaner-fish mutualism is remarkable both at the proximate and ultimate levels.

At the proximate level, one wonders how the fish – both clients and cleaners – know how to behave in different situations. At the ultimate (evolutionary) level, one wonders how natural selection favours the evolution of such cooperative behaviour and prevents it from disintegrating due to the temptation to cheat. Moreover, as humans, we cannot help viewing cleaner fish mutualism as civilised and morally uplifting.

Despite our obsession with ourselves, humans have long been concerned with the morality of animal behaviour, seen from our perspective, of course. Some have focussed on what they see as ruthless competition in the animal world and used it to either justify human behaviour or to admonish us to be better than animals. Others have marvelled at cooperation among animals and used it either to lament the lack of cooperation among humans or to encourage us to be at least as good as animals.

Thomas Henry Huxley (1825-1895), often dubbed “Darwin’s bulldog”, said:

“From the point of view of the moralist, the animal world is on about the same level as the gladiator’s show. The creatures are fairly well treated, and set to fight; whereby the strongest, the swiftest and the cunningest live to fight another day…the weakest and stupidest went to the wall, while the toughest and shrewdest, those who were best fitted to cope with their circumstances survived. Life was a free fight, and beyond the limited and temporary relations of the family, the Hobbesian war of each against all was the normal state of existence.”

On the other hand, the Russian prince, geologist, natural historian, and anarchist, Peter Kropotkin (1842-1921), saw cooperation everywhere. In his book-length critique of Huxley’s ‘gladiator’s view’ of animal life, Mutual Aid, he admonished

“Don’t compete!—competition is always injurious to the species, and you have plenty of resources to avoid it! That is the tendency of nature, not always realised in full, but always present. That is the watchword which comes to us from the bush, the forest, the river, the ocean. Therefore combine—practice mutual aid! That is the surest means of giving to each and to all the greatest safety, the best guarantee of existence and progress, bodily, intellectual, and moral.”

There has been a continuous conflict between the respective believers of competition and cooperation, even among biologists. This conflict is misguided because it is not a matter of competition versus cooperation, but of competition through cooperation. All would be well, therefore, if only the two camps of biologists cooperated!

Natural selection, survival of the fittest, struggle for existence – all phrases imply a competition between alternate forms for limited resources. Therefore, for an evolutionary biologist, competition is expected, and cooperation is indeed a paradox, whether in animals, humans or even plants or microbes. This does not mean that evolutionary biologists morally prefer competition over cooperation – nor that we privilege competition over cooperation as an explanatory principle for the patterns seen in nature. It does not even mean that competition is more common than cooperation. It simply means competition is easier to explain based on the theory of evolution by natural selection. On the other hand, cooperation, and especially altruism, requires special explanations.

Indeed, it is the need for special explanations that make cooperation and altruism most interesting and attracts so many of us to take on the challenge posed by nature. Contrary to the all too frequent complaint that we are obsessed with competition, those who assert that competition is expected and cooperation is a paradox are actually more interested in cooperation. Understanding the paradox of cooperation and altruism has been a major preoccupation of the branch of evolutionary biology that we call behavioural ecology.

The British biologist W.D. Hamilton provided one important solution to the paradox in 1964. Hamilton showed that cooperation, altruism and even self-sacrifice could be favoured by natural selection if the benefit of these acts is directed towards close genetic relatives. Hamilton’s solution, which has come to be known as ‘kin selection’, helps explain many instances of cooperation and altruism. However, there are far too many instances of cooperation between non-relatives, indeed, between different species, as we have seen in the cleaner-fish mutualism.

Seeing this gap in our theoretical framework, the American anthropologist and biologist Robert L. Trivers proposed the theory of reciprocal altruism in 1971. Trivers argued that natural selection could favour altruism if an act of altruism is returned in the future – i.e. if it is like a loan that is repaid. One can easily imagine a situation where both parties benefit if they help each other when one has plenty to give, the other is in need, and if the tables are turned often enough.

Thus the question is not so much whether reciprocal altruism will work in theory but whether it works in practice. Interestingly, cleaner-fish mutualism is one of the three examples that Trivers used to build his theory; the other two were warning cries (alarm calls) among birds and what Trivers simply called ‘human reciprocal altruism’.

With Trivers sowing the seeds of the interesting idea of reciprocal altruism, several researchers fleshed out the details. Some realised that the exact situation where one individual helps another today and the second individual is available and willing to help tomorrow when the first individual is in need might be a very rare occurrence. So they postulated the idea of indirect reciprocity: “I will help you because I have seen you help others, which gives me hope that you might help me someday when I am in need.”

Even this situation may be quite rare, so yet others have postulated the idea of generalised reciprocity: “I will help you; even though you have not helped me in the past, I am feeling good because someone has recently helped me.” Another idea is: “though I have not personally witnessed you helping others, I know that you have a reputation for helping others in need”.

The idea of reputation leads to the possibility, even necessity, that individuals eavesdrop on others and develop a system of image scoring, a kind of point system to keep track of the generosity of their neighbours.

The more complex the idea of reciprocal altruism became, the more attractive it seemed as a framework to understand human reciprocal altruism where, as Trivers said, “friendship, dislike, moralistic aggression, gratitude, sympathy, trust, suspicion, trustworthiness, aspects of guilt, and some forms of dishonesty and hypocrisy can be explained as important adaptations to regulate the altruistic system”.

However, the complications seem to make it unlikely to be applicable to animals. Keeping track of who helped whom, when and whether the help was returned and eavesdropping to assess the reputation of various individuals all seem to demand a highly developed cognitive ability, probably out of reach of most animals. Researchers have responded to this difficulty in two diametrically opposite ways.

Some have relentlessly pushed the boundaries of animal intelligence and cognition, and uncovered hitherto unsuspected levels of sophistication. Personally, I have much sympathy for this approach and have often endorsed it. The spirit behind this approach is reflected in the title of Frans de Waal’s book, Are We Smart Enough To Know How Smart Animals Are? True, de Waal’s book deals mainly with birds and mammals and especially primates, but others have made similar arguments about invertebrates, the most recent of which is Lars Chittka’s The Mind of a Bee.

On the other hand, Redouan Bshary has taken a radically different approach – one that may be more powerful and more widely applicable in the long run. Bshary says there can be strategic behaviour without cognition. Undaunted by the limited cognitive abilities expected of fishes, Bshary and colleagues have gone on to demonstrate that the strategic behaviour of cleaner fish and their clients is no less sophisticated, or perhaps more sophisticated, than even that of birds and primates.

In an early observational study, Bshary found considerable variation among the cleaners in their propensity to cheat, i.e., to eat the mucus or healthy tissue rather than the ectoparasites. Clients took the time to observe the quality of the service provided by individual cleaners to other clients and then approached the non-cheaters while avoiding the cheaters.

Thus the cleaners built up positive or negative reputations in the clients’ reckoning, who are said to be ‘image scoring’. The cleaners with a high propensity to cheat (low image score) employed an interesting counter-strategy. They behaved well, i.e. removed only ectoparasites and refrained from eating mucus when servicing small resident clients. Watching this good behaviour, the large, visiting clients approached them, and now the cleaners cheated on the big clients – a form of tactical deception.

Bshary argues that cleaners could learn the association between their behaviours and the resulting rewards and punishments without cognitively plotting a tactical deception strategy, as we humans or primates might be wont to do.

Bshary’s research team. L-R: Redouan Bshary, Nikki Raihani, the Lizard Island Research Station co-director Lyle Vail, Sharon Wismer and Ana Pinto. Photo: Redouan Bshary

Researchers often study reciprocal altruism using the tools of game theory, a mathematical modelling framework first developed in economics and profitably imported into evolutionary biology by John Maynard Smith and others. The best-known game in town is the so-called prisoners’ dilemma, in which both players are better off if they cooperate, but each is better off cheating if the other is going to cheat. Hence the dilemma. But sometimes, not knowing what the other player will do, both end up cheating.

However, mutual cooperation can become evolutionarily stable if the game is played repeatedly between the same players. For instance, if players cooperate in the first round no matter what the opponent does and then do whatever the opponent does in subsequent rounds, such a tit-for-tat strategy can lead to mutual cooperation.

Unfortunately, standard game theory and prisoners’ dilemma models don’t work too well for cleaner-fish mutualism. One problem is that many players are involved, and another is that, unlike in a standard prisoners’ dilemma, the client-cleaner fish game is not symmetrical. Here cleaners have greater opportunities to cheat than the clients, especially if the clients are non-predatory. Non-predatory clients are known to respond to cheating on the part of cleaners by attempting to punish.

As we already saw at the beginning of this article, the clients can take recourse to two forms of punishment: they can chase and harass the cheating cleaners or swim away and leave the cleaners with neither the less-preferred meal ectoparasites nor the more preferred meal of mucus. But do these punishments work? Do they result in better behaviour by the cleaners?  We need to know the answers to these questions before we can conclude that clients elicit cooperation from the cleaners by punishment.

To answer these questions, Bshary teamed up with Alexandra Grutter, a professor of integrative biology at Queensland University, Australia. Grutter had been studying the cleaner wrasse for even longer and was interested in more ecological questions before she collaborated with Bshary. Bshary and Grutter engineered a situation where they could make the clients appear to go away or chase a cheating cleaner and see whether this led to more cooperative behaviour by the cleaners.

To do so, they experimented with 24 cleaner fish caught from the coral reefs surrounding the island and acclimatised them to their laboratory conditions. Then they trained these fish to feed on prawn and fish flakes spread on a plexiglass plate.

Redouan Bshary and Alexandra Grutter, posing with their experimental cleaning wrasses. Photo: Redouan Bshary and Alexandra Grutter

The experiment consisted of three phases. In the first ‘test phase’, they offered pieces of prawn and fish flakes randomly placed in 1 X 1 cm cells in a 5 X 3 grid on the plexiglass. It became clear that the cleaners preferred prawns over fish flakes so that in the rest of the experiment, prawns stood in for mucus and fish flakes for ectoparasites.

In the second ‘learning phase’, the 24 fish were divided into three groups. The first control group of eight fish were allowed to feed as they pleased, so that they had no chance of learning that feeding on the preferred item (prawn) would have any negative consequences. In the second ‘client flees’ group, the plate was immediately removed if the fish ate the prawn but not if they ate the fish flakes, so that they could learn that the client would swim away if they consumed their preferred prey.

In the third ‘client chases’ group, the experimenters used the same plexiglass plate to chase the fish away if they ate prawn but not when they ate fish flakes, so that the cleaners could learn that they would be chased if they ate prawn.

In the third ‘foraging phase’ of the experiment, cleaners were allowed to eat whatever they wished, to see if their new preference had changed due to learning. Had the fish that had been chased or deprived of food if they ate prawn learned to avoid prawn and be content with eating fish flakes, any more than the control fish?

The answer was clear: the experience of being chased or deprived of food in just six trials had made the fish now eat their less preferred fish flakes, but there was no such effect on the foraging behaviour of the control group. We can conclude, therefore, that the swimming away or chasing response of the clients can induce cooperation by the cleaner fish.

A cleaner wrasse making a feeding choice from the plexiglass plate in the laboratory. Photo: Redouan Bshary

Bshary and Grutter set up another experiment to test whether clients indulge in image scoring and whether this benefits them. They set up a fish tank with mirrors so that bystander clients could watch the cleaners’ performance while servicing other clients. Their results showed that clients do watch cleaners in action and later spend more time next to well-behaved (cooperative) cleaners than misbehaving (cheating) cleaners. Even more interestingly, cleaners responded by behaving more cooperatively while being watched by bystanders.

My young colleague Hari Sridhar has the unusual habit of interviewing the authors of important papers and making them discuss the making of their papers, the back-chat as it were, bringing out many interesting anecdotes that tend to get left out in the scientific papers. Bshary has many interesting things to say about this work in one of Hari’s interviews. Bshary also gives a fascinating overview of his approach to the study of cleaner fish mutualism in another unusual interview format called ‘The Dissenter’, conducted by Ricardo Lopez in Portugal.

Bshary’s research strategy has been to make observations in the field, generate hypotheses about how the cleaners and clients might ensure stable cooperation, unmindful of the supposed cognitive demands of the hypothesised behavioural strategies, and then test the hypotheses in cleverly designed laboratory experiments. In a flurry of recent papers, Bshary and his colleagues have demonstrated that clients and cleaner fish match, and may even surpass, the abilities of primates in performing intricate strategic behaviours.

Using the relatively new theoretical tools of biological market theory and network theory, Bshary and his colleagues attempt to understand how these fish can pull off such complex behavioural strategies despite their admittedly limited cognitive capacities. This research has the potential of showing that highly developed brains with their advanced cognitive capabilities are unnecessary for performing complex behaviour.

More importantly, it has the potential to call into question the implicit assumption that birds and primates use their advanced cognitive capacities to shape their complex behaviour. Could it not be that birds, primates, including humans, also use the same strategies that fish use?

This line of reasoning may, in turn, call into question the widely accepted ‘social intelligence’ hypothesis. This hypothesis argues that the complex social environment primates live in and the need for them to adopt a strategic balance between cooperation and conflict were responsible for the evolution of their large brain and the resulting cognitive skills in primates and eventually in humans.

It is a sobering thought that studies of some lowly fish can challenge our cherished idea about the evolution of human intelligence. It is an even more sobering thought that we understand so little about the millions of species that inhabit, and for some, we may have to say soon, had inhabited our planet. Most sobering of all, I failed to find any research on cleaner fish mutualism carried out on the vast Indian seas.

Raghavendra Gadagkar is a Department of Science and Technology (DST) Year of Science Chair Professor at the Centre for Ecological Sciences at the Indian Institute of Science, Bengaluru.

Scroll To Top