Cleaner Wrasse Use Dropped Food to Probe How Mirrors Work

The mirror test for self-recognition has a straightforward logic: mark an animal where it cannot see the mark directly, show it a mirror, and watch whether it investigates the mark on its own body. Animals that do - humans, great apes, elephants, dolphins, magpies - are considered to show evidence of self-awareness. Cleaner wrasse have already passed that test in previous work by the Osaka Metropolitan University group. A new study from the same lab documents something more unusual: after passing the standard test, some wrasse began using the mirror in ways that suggest they were figuring out what mirrors actually are.

The behavior - picking up a small piece of shrimp from the tank floor, carrying it upward, and deliberately dropping it near the mirror while closely tracking its descent in the reflection - was observed in several fish after multiple days of mirror exposure. The researchers, led by Specially Appointed Researcher Shumpei Sogawa and Specially Appointed Professor Masanori Kohda at the Graduate School of Science, interpret this as contingency testing: using an external object to probe how the reflected world corresponds to reality, rather than using the animal's own body as the test object.

Contingency Testing and What It Implies

Contingency testing - watching how an object's movement in the mirror corresponds to its real-world movement - has previously been documented in dolphins, which release bubbles and observe them rise in mirror images, and in manta rays. In each case, the animal appears to be investigating the mirror itself as an object with predictable properties rather than confusing its reflection for another individual of the same species. This is distinct from simple social confusion (attacking or displaying at the mirror) and distinct from the mark-directed behavior of the standard mirror test: it is an active investigation of how the mirror works.

The significance lies in what the behavior implies about the cognitive processes underlying it. For a fish to pick up food, carry it to a specific location, release it, and then track its descent while observing how the reflected image corresponds to the object's actual movement requires holding multiple representations simultaneously: the fish's knowledge of its own position, the object's position and trajectory, and the mirror image's position and trajectory. That is not a behavior easily explained by simple conditioning or reflexive responses.

The Methodological Twist That Accelerated Recognition

The study also produced an unexpected finding about the standard mirror test procedure. In conventional experiments with cleaner wrasse, the protocol has been to expose fish to a mirror for several days first - allowing them to habituate and stop treating the reflection as another fish - before applying the mark. Recognition behavior, when it appears, typically emerges four to six days after marking.

The Osaka team reversed the sequence: they marked the fish first, then introduced the mirror for the first time. Under this reversed protocol, mark-directed scraping behavior appeared on average within 82 minutes of first mirror exposure - dramatically faster than the four-to-six-day timeline previously reported.

Dr. Sogawa's interpretation is that fish marked before seeing a mirror are already aware that something unusual exists on their body - they just cannot see it. When the mirror appears and provides visual information that matches that bodily expectation, the fish immediately connects the two representations, and scraping behavior follows quickly. The previous protocol's multi-day delay may have reflected the time needed to recognize one's own reflection as self-relevant rather than as another fish, not the time needed to then act on what the reflection shows.

What This Means for Understanding Animal Self-Awareness

The mirror test has been criticized as an imperfect instrument for measuring self-awareness. It may undercount the phenomenon because it privileges visual self-recognition in species for whom vision is the dominant sense - a bias that disadvantages species relying heavily on olfaction, echolocation, or other modalities. It may also set an inappropriately high threshold for recognition behavior, since passing the standard test requires not just self-awareness but the motivation to groom a mark in a species that typically grooms itself in the first place.

The wrasse data extends the test in a different direction: beyond the mark test, toward understanding what animals can learn about mirrors as objects. Contingency testing represents a more sophisticated level of mirror engagement than simple self-recognition, and its appearance in fish - which diverged from the lineage leading to mammals and birds more than 400 million years ago - suggests that the cognitive prerequisites for this behavior are not restricted to the vertebrate lineages where it has previously been documented.

Dr. Sogawa suggested that self-awareness may not have evolved exclusively in the limited number of species that have passed the mirror test, and that mirror self-recognition may be more widespread than current evidence indicates. Professor Kohda framed the broader implications as extending to evolutionary theory, concepts of self, animal welfare, and even AI research - the last because understanding what computational architectures support self-referential processing is an open question in AI cognition that benefits from knowing which biological systems achieve it.

Limitations

The study involves a small number of fish in controlled tank conditions, and the contingency-testing behavior was observed in some but not all individuals. Inter-individual variation in mirror behavior is common in cleaner wrasse research and is itself an interesting phenomenon that the field has not fully explained. Whether the behavior observed in tanks reflects cognitive capacities that cleaner wrasse routinely deploy in the wild - where mirrors do not exist and the relevant cognitive demands are different - cannot be determined from this study design.

The study was published by Osaka Metropolitan University's Graduate School of Science. The research team plans to extend their investigation of self-awareness to a wider range of species, including invertebrates.

Cleaner Wrasse and the Biology of Self-Awareness

Cleaner wrasse occupy an unusual ecological niche that may predispose them to sophisticated social cognition. They operate cleaning stations on coral reefs where they remove parasites, dead tissue, and food debris from client fish, including species that would normally prey on them. Successfully running a cleaning station requires tracking individual client fish across repeated visits, managing social relationships with clients of varying temperament, and distinguishing between contexts where biting a client is acceptable and where it is not. This ecology places demands on social memory and behavioral flexibility that are uncommon among fish.

Whether these ecological demands have driven the evolution of self-awareness specifically, or whether self-awareness is a more general feature of fish cognition that the cleaning ecology simply makes more visible in mirror tests, is not resolved by this study. What the Osaka team's cumulative research program has established is that cleaner wrasse exhibit a behavioral flexibility in mirror-related tasks - from mark-directed scraping to contingency testing using external objects - that overlaps substantially with the behavioral markers used to infer self-awareness in cognitively sophisticated mammals. The evolutionary distance between fish and mammals makes this either a remarkable case of convergent cognitive evolution or a sign that the cognitive requirements for self-referential processing are more modest than has been assumed.