The Kya Project
28
Apr 2026
Read More
.svg)
In a study tracking 411 individually identified wild sulphur-crested cockatoos across three Sydney roost sites, researchers from the Max Planck Institute of Animal Behavior and Australian National University found something unexpected: parrots don't use a single social strategy. They use two completely different ones depending on who they're dealing with. You can read the complete study in the original paper.
Most research on animal social hierarchies assumes a relatively stable, closed group. The same individuals interact repeatedly, build a history, and over time establish a clear sense of who outranks whom. Conflict becomes less frequent because everyone already knows where they stand.
Parrots don't live like that.
Wild sulphur-crested cockatoos sleep in communal roosts of up to 200 birds, then leave each morning in constantly shifting subgroups that can be anywhere from 2 to 500 individuals. On any given day, a bird might forage alongside its closest companions, birds from a neighbouring roost it recognises but knows less well, and complete strangers passing through. The flock fragments and reforms constantly — what researchers call fission-fusion dynamics.
This creates a problem that no amount of memory alone can solve. You cannot maintain detailed social records of every individual you might ever encounter. So how do cockatoos — and by extension, most parrots — actually navigate that complexity without constant costly conflict?
That is what this 2025 study set out to answer. And what it found changes how we should think about parrot social behaviour, including what happens inside multi-bird households.
The study was conducted with wild sulphur-crested cockatoos in urban Sydney — a species that lives in year-round communal roosts likely spanning decades, with individuals foraging across a roughly 3km radius and regularly encountering birds from neighbouring roost communities.
Researchers marked 411 individual birds using unique combinations of non-toxic coloured dye, allowing each bird to be identified in the field. They also trained birds to step voluntarily onto a scale in exchange for a food reward, collecting weight data as a measure of body size and physical capacity. Social associations were recorded through regular presence scans, and every aggressive interaction — who initiated it, who responded, and how — was logged.
The team then divided each bird's interactions into two categories:
This separation is what makes the study's design so useful. Most research treats all interactions within a species as comparable. This study asked whether the same bird makes decisions differently depending on how much social information it has about the individual in front of it.
The results were clear.
With familiar birds — those a cockatoo knew from repeated encounters at its home roost — social decisions were driven by dominance rank. Birds initiated aggression toward those close in rank, avoided those well above them, and responded to challenges based on where the initiator sat in the hierarchy. This is memory-based social cognition: the bird knows this individual, knows their history together, and uses that knowledge to decide whether to engage.
With unfamiliar birds, the system switched entirely. In the absence of social memory, cockatoos fell back on a physical shortcut: body weight. Interactions were directed toward birds of similar size, and the decision to escalate or retreat was governed by how comparable the two birds appeared physically.
With birds they know, cockatoos use memory. With strangers, they read the body.
This two-track strategy allows cockatoos to navigate social complexity at scale. They don't need a complete social record of every bird they might encounter. They have a fallback system for strangers that doesn't require memory at all — just a rapid physical assessment.
What makes this especially striking is that sulphur-crested cockatoos have no visible plumage differences between individuals that would signal rank at a glance. Unlike house sparrows, whose bib size advertises dominance, or great tits with their variable breast stripe, every cockatoo looks essentially the same. So they cannot read a stranger's status from appearance alone. Weight — something they can assess through direct observation — becomes the proxy instead.
The study's definition of familiarity is worth sitting with, because it reframes a concept owners often use loosely.
Familiarity here is not just recognition. It is the accumulation of enough repeated social interactions to build a working model of how that individual behaves — how they respond under different conditions, where they sit in relation to you, whether they are predictable. A bird that has seen another bird from across a room is not familiar with it in this sense. Familiarity is built through social history.
This matters because it explains why introductions between parrots can feel so unstable even when the birds are not overtly aggressive. The established bird is not simply being territorial. It is operating without the social information it normally uses to calibrate its behaviour. It has switched into stranger-assessment mode — which is blunter, more cautious, and built around physical comparison rather than nuanced relational history.
The new bird is not yet a flock member in any cognitively meaningful sense. It is an unknown quantity.
The study also confirmed something about how dominance is structured in parrot societies more broadly.
The hierarchies the researchers documented were real and measurable — but they were also shallow. Not every bird had a clearly distinct rank from every other bird. There was meaningful ordering, but also a lot of near-equivalence, particularly among adults of similar age and sex.
Rank was influenced by:
Crucially, the hierarchies extended beyond the home roost. Birds maintained rank relationships with individuals from neighbouring roost communities — birds they encountered less often but still knew. Social memory in cockatoos is not confined to their immediate daily companions. It scales outward through a wider network, with familiarity and interaction frequency determining how precise that knowledge is.
For anyone living with more than one parrot, this research offers a more precise framework than the usual advice around introductions.
When a new bird arrives in a home where a bird already lives, the established bird faces exactly the scenario this study describes: an unfamiliar individual with unknown social standing. Its default system — the one built on memory and nuanced rank assessment — has nothing to work with. So it switches to the stranger-assessment mode: reading physical cues, monitoring proximity, and treating the new bird as an unknown competitive variable rather than a known social partner.
This is not aggression in the simple sense. It is a different cognitive mode. And it explains several things that confuse owners during introductions:
The goal of a slow introduction is not just to reduce immediate conflict. It is to give both birds enough shared social history that they can move from stranger-assessment mode into the more nuanced, memory-based system that stable flock relationships depend on.
Introductions take time not because parrots are stubborn — but because social familiarity is something that has to be built, not assumed.
The deeper implication of this study is that parrot social behaviour is more context-dependent than it often appears.
The same bird, interacting with different individuals, is running fundamentally different cognitive processes. With a long-known companion, it draws on a rich relational history. With a stranger, it is essentially starting from scratch, reading what it can from physical cues alone.
Owners sometimes interpret changes in a bird's behaviour as personality shifts or mood fluctuations. Often what they are seeing is the bird responding to a change in social context — encountering someone unfamiliar, being placed in a new group, or having the composition of its social world shift in ways that require recalibration.
Understanding that parrots carry this dual system — one for the known, one for the unknown — gives a more accurate lens for reading what is actually happening in those moments.
This study is significant for what it demonstrates about parrot cognition as much as for what it reveals about parrot social life. Maintaining two parallel assessment strategies, switching fluidly between them depending on available information, and sustaining stable hierarchies across a socially complex fission-fusion society — this is cognitively demanding in ways that mirror the social intelligence seen in primates and cetaceans.
Parrots are not just reacting to whoever is in front of them. They are drawing on social memory, running assessments, and making decisions calibrated to how much they know.
That applies in the wild, across hundreds of birds in shifting urban flocks. And it applies in a living room, between two birds who have not yet built enough history to trust what they know about each other.
This 2025 study by Penndorf, Farine, Martin, and Aplin tracked 411 individually marked wild sulphur-crested cockatoos across three neighbouring roost communities in Sydney, recording social associations, individual weights, and aggressive interactions across two observation periods. The researchers tested whether cockatoos use different social decision-making strategies depending on their familiarity with the individual they are interacting with.
They found that cockatoos operate a two-track system: using dominance rank — built from memory of repeated interactions — to navigate encounters with familiar birds, and switching to body weight as a physical proxy when interacting with less familiar individuals. The hierarchies themselves were stable and extended across roost boundaries, incorporating birds from neighbouring communities the focal birds knew but encountered less frequently.
Stay updated with our charity’s news and see the impact we’re making changes together.
Copyright © The Kya Project
.jpg)
