An empirical study of large, naturally occurring starling flocks: a benchmark in collective animal behaviour
Bird flocking is a striking example of collective animal behaviour. A vivid illustration of this phenomenon is provided by the aerial display of vast flocks of starlings gathering at dusk over the roost and swirling with extraordinary spatial coherence. Both the evolutionary justification and the mechanistic laws of flocking are poorly understood, arguably because of a lack of data on large flocks. Here, we report a quantitative study of aerial display. We measured the individual three-dimensional positions in compact flocks of up to 2700 birds. We investigated the main features of the flock as a whole - shape, movement, density and structure - and discuss these as emergent attributes of the grouping phenomenon. We find that flocks are relatively thin, with variable sizes, but constant proportions. They tend to slide parallel to the ground and, during turns, their orientation changes with respect to the direction of motion. Individual birds keep a minimum distance from each other that is comparable to their wingspan. The density within the aggregations is non-homogeneous, as birds are packed more tightly at the border compared to the centre of the flock. These results constitute the first set of large-scale data on three-dimensional animal aggregations. Current models and theories of collective animal behaviour can now be tested against these results.
💡 Research Summary
The paper presents the first large‑scale quantitative analysis of starling murmurations, providing a benchmark dataset that can be used to test and refine theories of collective animal behaviour. Using a synchronized multi‑camera system, the authors reconstructed the three‑dimensional positions of every individual in flocks ranging from a few hundred up to 2 700 birds. This unprecedented resolution allowed the authors to examine the flock as a whole—its shape, motion, density, and internal structure—and to infer the emergent rules that govern the grouping phenomenon.
Key morphological findings are that flocks are relatively thin, roughly cylindrical or flattened ellipsoidal bodies whose aspect ratios (width : depth : height) remain constant (≈3 : 3 : 1) even as absolute dimensions vary from 10 m to 30 m. This suggests that individuals adjust their spacing to preserve a global geometric proportion, a constraint not captured by many existing models.
Kinematically, flocks travel primarily parallel to the ground at an average speed of about 12 m s⁻¹. During turns the entire group rotates around a common axis while the direction of motion changes relative to the flock’s orientation. This coordinated rotation implies that information about curvature propagates through the group faster than simple velocity‑matching rules would predict, hinting at a higher‑order interaction such as visual torque transmission among neighbours.
The spatial organization within the flock is markedly non‑homogeneous. Birds maintain a minimum inter‑individual distance comparable to a single wingspan (≈0.4 m), indicating a hard personal‑space constraint that prevents collisions. More intriguingly, density is higher at the periphery than at the centre; birds at the edge are packed roughly 10 % more tightly. This “border compression” could serve defensive functions (e.g., reducing predator access) or arise from visual constraints that make individuals at the edge more sensitive to neighbours.
Statistical analysis of nearest‑neighbour distances shows that each bird interacts with about 6–7 closest conspecifics, consistent with the classic “seven‑plus‑or‑minus‑two” rule often assumed in models. However, the interaction range is not a fixed metric distance but rather a function of angular visibility and relative position, reinforcing the importance of anisotropic visual fields in starling flocks.
When the authors compare their empirical results with prevailing theoretical frameworks—such as the Boids model, visual‑interaction models, and self‑organized criticality approaches—they find partial agreement. Existing models reproduce overall flock cohesion and speed alignment, but they fail to generate the observed peripheral densification, the constant aspect‑ratio maintenance, and the precise turning dynamics. Consequently, the authors argue that additional rules—e.g., enhanced edge‑agent coupling, torque‑based orientation updates, or density‑dependent repulsion—must be incorporated to achieve realistic simulations.
In summary, this study delivers a comprehensive dataset of three‑dimensional starling murmurations, quantifies fundamental geometric and dynamical properties of large flocks, and highlights specific phenomena that challenge current collective‑behaviour models. By establishing concrete, measurable benchmarks, the work paves the way for a new generation of data‑driven, biologically grounded theories of animal group motion.