Abstract Protogynous sex change, where individuals first function as females and later as males, is a key life-history strategy among polygynous reef fishes. In haremic systems, sex change is typically socially regulated, with dominants suppressing subordinates’ sex change through aggression. Females within a harem form a size-based hierarchy that can remain stable in most species through the threat of eviction. We studied a different situation in the cleaner wrasse Labroides dimidiatus , where larger females have incomplete control, as they spend most of their time alone at their own cleaning territory. We tracked over 400 individuals for 12 months, recording growth, behavior, social organization, and sex change. We confirmed earlier reports that both sexes direct aggression primarily at those ranked immediately below them. However, we observed 30 cases where smaller females outgrew larger ones, revealing hierarchy instability. Of 42 sex change events, 43% occurred in presence of the male, and half of these ‘early’ sex changers were not the largest female, but individuals overlooked by the male. Fast growth relative to harem-mates and harem switching increased the likelihood of sex change. Local population densities also influenced growth and sex change, with individuals in high-density demes growing faster and changing sex at larger sizes. Our findings reveal flexible sex change dynamics in a system with incomplete social dominance. Such incomplete control and observations that becoming male confers both higher reproductive success and survival highlight the need to expand game-theoretical and life-history frameworks to encompass such strategic flexibility. Lay summary Dominant cleaner wrasse cannot fully control subordinates as individuals occupy distinct core areas. Tracking 400 fish for a year, we found that smaller females could outgrow initially larger ones, and ‘early’ sex change despite a larger male. Fast growth and harem switching increased the chances of becoming male. Population density also shaped these strategies. Our findings reveal flexible sex change dynamics in a system where becoming male confers both higher reproductive success and survival.

Abstract In many polygynous species, males face stronger intrasexual competition, higher energetic demands, and lower survival than females, especially under resource limitation or environmental stress. Such sex-specific vulnerabilities are expected to intensify with climate change. Yet, in sequentially hermaphroditic systems, where individuals change sex during their lifetime, how sex and sex change shape survival remains largely unexplored. We studied sex-specific survival and growth in the haremic protogynous cleaner wrasse Labroides dimidiatus across eight reefs around Lizard Island, Great Barrier Reef. We tracked a total of 731 adult fish (individually recognizable through marking or idiosyncratic color patterns) over two years. This period included the 2024 El Niño–Southern Oscillation (ENSO), which caused a temporary 1-degree increase in water temperature, severe coral bleaching, and coral mortality at Lizard Island. Contrary to expectations from dioecious systems, terminal-phase males exhibited higher survival than initial-phase females under both normal and in particular ENSO conditions. While male mortality was not affected, female mortality more than doubled during the event, indicating greater physiological or energetic vulnerability. A partial explanation for the overall higher female mortality is their generally faster growth rate, which declined in both sexes during the ENSO event. Our findings challenge existing assumptions of male-biased mortality in polygynous species and highlight that sex and sex change fundamentally shape demographic responses to climate extremes.

Collective defense of territories and resources is a key benefit of group living and is often modeled as a public goods game. Although larger animal groups have the potential to mobilize greater fighting power in between-group competition, this advantage may be constrained by coordination. Our review of quantitative studies shows that, independent of a population's average group size, larger groups usually win, with few exceptions. We examined one such case in vervet monkeys. As expected, group size was positively correlated with group spatial spread, and more dispersed groups were less likely to win encounters, implicating coordination problems. We then developed a mathematical model showing that, as coordination costs rise with group size due to spatial spread, they can cancel the usual "strength in numbers" advantage of larger groups in between-group conflicts. Our work highlights how between-group competition provides a natural setting for integrating ecological constraints into models of N-player cooperation.

Despite the astonishing diversity of life, evolutionary biologists have documented numerous instances where species converge in how they behave, look and function. Given the importance of happenstance in evolution, it is often assumed that mechanisms driving independent evolution of similar phenotypes (convergence) are distinct; yet recent discoveries suggest that conserved genomic mechanisms can underlie convergent traits. Here, we generate forebrain transcriptomes from six sympatric Labridae wrasses that vary in mutualistic cleaning behaviour and apply differential gene expression, co-expression and phylogenetic comparative analyses to address three questions. First, do neurotranscriptomes vary among species with distinct cleaning behaviour? Second, do species that have independently evolved facultative cleaning share neurotranscriptomic patterns? Third, are transcriptomic signatures linked to facultative cleaning also expressed in obligate cleaners? We identify shared neurotranscriptomic patterns in the repeated evolution of facultative cleaning, including 25 novel candidate genes. We then find that transcriptomic patterns associated with facultative cleaning are shared in our focal obligate cleaner species, with more genes than expected by chance with concordant expression patterns in obligate and facultative cleaners. Our results illuminate the neuromolecular basis of cooperative behaviour and demonstrate the predictive potential of comparative transcriptomics to unravel mechanistic underpinnings of the repeated evolution of complex phenotypes.

The evolution of cooperative behaviour among unrelated individuals is still incompletely understood. One example is the provision of costly services by males in primate societies, which is difficult to explain by traditional reciprocity-based models. Here, we investigated group progressions in wild vervet monkeys (Chlorocebus pygerythrus) to test two main hypotheses for male-biased positioning in potentially risky contexts: paternal care and reputation-based partner choice. Over more than 2 years, we recorded the order and timing of individual crossings in five neighbouring groups of vervet monkeys across two terrain types differing in presumed risk (rivers and dirt roads). Adult males, particularly dominant males, were more likely to occupy leading positions than females or juveniles, and this tendency was stronger during river crossings. When leading, males often crossed alone, whereas non-leading males usually crossed within subgroups. Leadership was most pronounced in dominant males without likely offspring. Males who did not lead did not preferentially cross with infants based on their paternal status, and higher leading rates did not predict subsequent mating success or short-term changes in dominance rank. These findings provide no support for paternal care or direct mating benefits. Instead, male leadership in these contexts likely reflects a combination of ecological conditions, and dominance- and paternity-related differences in behaviour. We conclude that male-biased positioning in group movement is context-dependent and may arise from multiple, potentially overlapping processes rather than a single cooperative function. Our results highlight the role of dominance and reproductive status in shaping male service provisioning and underscore its potential flexible function in primate social organisation.

An important challenge for individuals of group-living species is to build cooperative relationships with new partners. One famous strategy for doing so is 'raise the stakes', proposing low initial investments that increase if the partner matches these investments in a series of exchanges. Contrarily, the 'all in' strategy predicts high initial investment that may be maintained, but downregulated if not fully matched. We tested these predictions on grooming exchanges between wild male-female vervet monkeys, a species with regular male dispersal. Contrary to model predictions, we found uneven initial investment between novel partners that reached near reciprocity after approximately 6 months. To identify which sex altered investment, we examined grooming durations and frequencies of novel and established partners. Females showed consistently high initial investment and gradually reduced grooming over the first 6 months. Males did not alter grooming duration but had higher grooming frequency after a year of residency. Eventually, grooming exchanges nearly evened out. Female behaviour aligns more closely with the 'all in' strategy, which suggests competition among female groups over dispersing males. Male behaviour partly fits the 'raise the stakes' hypothesis. The study highlights the complexity of real-life cooperation, emphasizing the need to explicitly incorporate life history parameters to understand cooperative strategies.

The observed difference in relative brain size between endotherms and ectotherms raises questions about potential resulting disparities in brain function between these two groups. Until recently, no clear cognitive advantage was found in endotherms, with ectotherms occasionally even outperforming them in seemingly complex tasks. However, recent research on working memory-a core executive function-in a teleost fish species suggests that cognitive differences may lie in more fundamental processes. Here, we develop two working hypotheses that arose from this finding. First, the apparent absence of working memory in a fish, and possibly other ectotherms, may stem from their inability to voluntarily control their attentional focus. Instead, the environment would drive, through other cognitive processes, changes in that focus. In 2011, in the dichotomic vision of Kahneman consisting of automatic System 1 and voluntary System 2, fish could rely only on System 1. We call this the Lack of Attentional Control hypothesis. Second, to explain why smaller-brained species may nevertheless outperform larger-brained species in some cognitive tasks, we propose the Cognition-Opportunities-Needs framework, which posits that cognitive abilities and learning opportunities provide non-mutually exclusive mechanisms for meeting ecological demands. Although these hypotheses require extensive empirical validation, they represent a first step toward a comprehensive theoretical perspective on cognitive diversity and evolution across species and major vertebrate clades.

Summary Interactions between cleaner fish Labroides dimidiatus and client fish, from which cleaners remove ectoparasites and mucus, represent a textbook example of mutualism involving sophisticated strategic decision-making. However, cleaners must also face intraspecific social challenges within a size-based hierarchy, where the largest females may eventually change sex and become males with higher reproductive rates. Following 540 individuals over 11 months, we found that, contrary to expectations, slow-growing females spent more time cleaning and cheated more frequently, without causing more negative client responses than fast-growing females did. Instead, variation in growth was best explained by social factors: fast-growing individuals experienced reduced social control, while slow growers spent more time in proximity to dominant individuals. As there was no evidence that spawning activity affected growth patterns, it appears that fast growth as a viable strategy for becoming a male largely depends on the lack of control by dominants.

There is tremendous taxonomic variation in the size, shape and structure of vertebrate brains. While many studies use cross-species comparisons to aim at identifying the ecological factors (social and environmental) that explain what cognitive advantages larger brains confer, a more fundamental divide exists between endotherm and ectotherm vertebrates. Relative to body size, ectotherms have ten times smaller brains than endotherms. Existing ecological hypotheses cannot explain this divide, as some endotherm species with relatively simple social organization and diets still possess larger brains. Furthermore, research demonstrates that at least fishes possess a cognitive 'toolkit' equivalent to that of many endotherms. This is the fish challenge to vertebrate cognitive evolution. We review hypothesized consequences of brain size differences to propose two non-exclusive solutions. First, the fish brain retains modularity, but it is less efficient in problem-solving than an endothermic brain with a more domain-general organization. Second, brain size variation can be more effectively explained by sensory-motor skills and their integration rather than by cognitive processes. In that case, understanding brain size would require highlighting and emphasizing these aspects when broadly defining animal cognition. Specifically, it would be fitting to amend the classic definition of animal cognition to refer to how animals take in and process sensory information before deciding how to act on it using motor competencies.This article is part of the Theo Murphy meeting issue 'Selection shapes diverse animal minds'.

How might brain functioning differ between endotherm and ectotherm vertebrates? Recent results suggest that ectotherms lack proper working memory, which could reflect a lack of attentional control. Ectotherms may nevertheless excel in cognitive tasks if their ecological needs and learning opportunities compensate for their lower computing power.

Working memory (WM), an attention-based short-term storage system responsible for the manipulation and integration of past knowledge with present information for goal-directed behavior, is a key executive function and a principal predictor of general intelligence. Despite its importance, WM has not been a major research topic in animal behavior. Here, we first summarize key ideas related to WM from the social sciences for interested colleagues. Given that past methodological inconsistencies have led to mixed results and conclusions across various species, we then designed experiments that incorporate the critical components of WM, facilitating cross-species comparisons and accounting for potential ecological influences. We present such experiments on WM in an ectothermic vertebrate, the cleaner wrasse (Labroides dimidiatus), which faces environmental challenges potentially requiring complex cognitive adaptations. Overcoming several experimental challenges, we consistently obtained negative results across multiple experimental paradigms. As our experiments were specifically designed to test WM, our negative results call into question previous studies in other fish species that provide evidence for WM using different paradigms. More specific tests for WM should be developed to confirm the presence or absence of this executive function in other ectotherm vertebrates. The absence of WM may be a key factor underlying the significant encephalization gap between ectotherm and endotherm vertebrate species.

The evolution of cooperative behaviour among unrelated individuals is still incompletely understood. One example is the reported service provisioning by males in primate societies, which is difficult to explain by traditional reciprocity-based models. In this study, we investigated group progressions and sex roles in wild vervet monkeys ( Chlorocebus pygerythrus ) to test two hypotheses for the observed male bias in cooperative and protective behaviours: paternal care versus reputation-based partner choice. Over more than two years, we recorded the order and timing of individual progressions in five neighbouring groups of vervet monkeys across two terrain types differing in risk (rivers and roads). Males, particularly dominant males, were more likely to lead than females or juveniles, and their probability of leading increased in high-risk terrain. When leading, males often crossed alone, thereby assuming higher predation risk, whereas when not leading, males usually crossed within subgroups. Dominant males who had no offspring yet were most likely to lead during risky progressions. Males who did not lead did not preferentially cross with infants based on their paternal status, refuting the paternal care hypothesis. Instead, male leadership in risky contexts appears to represent a cooperative service. While some of these findings are consistent with the reputation-based partner choice model, we found no clear evidence that such services translated into direct mating advantages. Overall, our results highlight the role of dominance and reproductive status in shaping male service provisioning and underscore its potential adaptable function in primate social organisation. ### Competing Interest Statement The authors have declared no competing interest. Swiss National Science Foundation, https://ror.org/00yjd3n13, 310030_197884, PP00P3_198913 University of Lausanne, https://ror.org/019whta54, ProFemmes European Research Council, https://ror.org/0472cxd90, 949379

Self-awareness in animals is often documented by showing evidence for mirror self-recognition (MSR), which is confirmed by the mirror mark-test. The classic assumption about self-awareness was that it is a complex cognitive process, restricted at best to a few large-brained species. Indeed, early MSR research yielded positive results only in great apes, elephants, dolphins and magpies, while most endotherm species failed. However, recent detailed proof of MSR based on a mental representation of self, i.e. private self-awareness (PrSA), in cleaner wrasse Labroides dimidiatus, a small-brained ectotherm fish, indicates that the origin and cognitive complexity of self-awareness must be reconsidered. Here, we first recapitulate key concepts on the evolution of self-awareness, and then summarize the evidence that cleaner wrasse exhibit PrSA. We propose that the many negative MSR results are potentially false-negatives, and that self-awareness does not require a large brain. We posit a new hypothesis: self-awareness was already present in the early shared ancestors of modern vertebrates.This article is part of the theme issue 'Evolutionary functions of consciousness'.

Supply and demand affect the values of goods exchanged in cooperative trades where high demand typically leads to a higher price. An exception has been described in the marine cleaning mutualism involving the cleaner fish Labroides dimidiatus and its variety of ‘client’ coral reef fishes. Cleaner fish feed on clients’ ectoparasites but prefer eating clients’ mucus instead, which constitutes cheating. Here, we provide field observations, followed by a set of laboratory experiments with real clients and Plexiglas feeding plates as surrogates for clients. In the field and in three experiments with real clients, we found that satiated cleaner fish were more cooperative, even though low hunger levels should make them less dependent on cleaning interactions. Similarly, the more abstract version of the experiments using Plexiglas plates offering two food types mimicking client ectoparasites and mucus showed that satiation led cleaner fish to feed more against their preferences – an indicator of cooperative behaviour. However, this outcome occurred only if the temptation to eat the preferred food was low. When the temptation to cheat was high, cleaners did so. We provide further general support to these findings with a game-theoretic model. Many mutualisms involve food as a commodity. Thus, identifying foraging decision rules will enhance our understanding of how individuals adjust to real-time market conditions rather than playing evolved strategies adapted to the average market conditions.

Various concepts have been developed to explain cooperation among unrelated individuals, but linking proposed strategies to empirical evidence remains a challenge, especially for variable reciprocal investments. One famous strategy is "raise-the-stakes", proposing low initial investments but a steady increase if the partner matches the other's investments or increases their own investments in a series of exchanges. Contrarily, the "making-friends" strategy predicts high investment between strangers that may either be kept up or downregulated if not fully matched. We tested these predictions on grooming exchanges between wild male-female vervet monkeys, a species with regular male dispersal. Contrary to both predictions, females had a consistently high initial investment and groomed immigrants for longer than was reciprocated, showing no adaptation to the formation of debts. However, females gradually reduced grooming over the first six months. Males, on the other hand, did not alter grooming investments over time. Thus, grooming exchanges eventually almost evened out. We present various potential explanations for these results, which are overall more consistent with the "making friends" strategy. Most importantly, this study highlights the complexity of real-life cooperation, emphasizing the need for models that explicitly incorporate life history parameters into cooperative strategies. ### Competing Interest Statement The authors have declared no competing interest.

Ectotherms, particularly fish, challenge traditional brain evolution theories by exhibiting advanced cognitive abilities despite their smaller brains. While the social brain hypothesis may apply within clades, sensory-motor systems likely explain the brain size differences between average-brained ectotherms and endotherms. Evolved complex sensory-motor systems suggest that brain evolution models should expand to include sensory and motor systems, beyond cognitive processes alone.

INTRODUCTION:The factors shaping vertebrate brain evolution and cognition are broadly categorized as being either social or environmental. Yet, their relative importance is debated, partly due to the limitations associated with standard interspecific evolutionary comparisons. Here, we adopt a complementary strategy leveraging within-population variation in fish brain size to ask how variation in social and environmental factors correlates with individual brain size. METHODS:We investigated how overall brain size and brain part sizes varied between demes of the same population in the coral reef-associated batu coris Coris batuensis. This species is ideal for our approach because its local population densities are dissociated from both interspecific densities and habitat complexity. RESULTS:We found that individuals from demes with higher population densities possess larger overall brain volumes than those from lower population density environments, caused by an enlargement of all five main brain regions. Brain anatomical measures show no correlation with interspecific density or habitat complexity. CONCLUSION:Our results suggest that variation in intraspecific social challenges is selected on individual batu coris brain size, either through phenotypic plasticity, differential survival, or habitat choice. These results conform with a broader version of the social brain hypothesis, emphasizing the importance of the entire brain over specific regions like the neocortex in mammals or the telencephalon in fishes.

Whether animals are self-aware has important implications for our approaches to both animal cognition and animal welfare. A landmark moment in animal cognition research was when great apes passed the mark-test and demonstrated mirror self-recognition (MSR). Animals that pass the mark-test are capable of visually self-recognising and considered to be self-aware. Other taxa, including a fish, the cleaner wrasse (cleaner fish: Labroides dimidiatus) have also now passed the mark-test, forcing a rethink of the mental and neurological requirements for MSR. Previous research has largely focused on which species can pass the mark-test, rather than the processes underlying MSR. Here, we marked mirror-naïve cleaner fish with an ecologically relevant mark resembling an ectoparasite and then undertook detailed behavioural observations after exposure to a mirror. We found that cleaner fish achieve MSR rapidly, implying self-awareness prior to mirror exposure. By observing the exact timing of MSR in individuals, we could also report previously undocumented differences in pre- and post-MSR behaviours, including post-MSR exploratory behaviour of the mirror’s reflective properties. We find remarkable parallels between the processing of MSR in humans and cleaner fish, suggesting that some aspects of self-awareness are conserved across animal taxa.