An experiment about empty space
The
dandelion
and the
ring in the air
The seed’s pappus is barely there: 91.6% of its enclosing disk is empty projected area. Flow through those gaps maintains a stable, separated vortex ring that slows the fall.
91.6% is the empty area seen in projection, divided by the plan area of the enclosing disk. It is not the percentage of air by volume.
Follow the airVertical wind tunnel
interactive diagramMeasured pappus: 91.6% empty projected area. In the experiment, flow through the filaments maintained a stable, separated ring.
The 0%, 33%, 55%, 75% and 89% stops come from the study’s porous disks. The ring’s drawn position between the endpoints shows the trend and does not mark a measured threshold.
How the pappus works
The gaps are part of the mechanism.
Air passes between the filaments. A region of slowed air forms around each one, and neighbouring filaments sit close enough for those regions to interact. The spread-out crown therefore behaves as a coherent aerodynamic object.
At filament scale, the measured Reynolds number was 0.422. The study estimates a mean spacing of roughly 41 filament diameters; the influence of neighbouring filaments might only become negligible beyond about 47.
- Filaments
- ≈100 95% CI: 95–106
- Pappus diameter
- 13.8 mm 95% CI: 13.2–14.3
- Terminal speed
- 39.1 cm/s 95% CI: 34.9–43.0
The speed was measured in still air across ten samples. It does not predict how far a seed will travel in wind.
The same measured force
The same force, far less material.
By the definition used in the study’s comparison, the pappus and the solid reference disk produce the same total drag at the same speed. The disk’s calculated diameter is 8.6 mm. Substituting the authors’ values in SI units gives an area of 57.7 mm²; the filaments add up to 12.6 mm².
Where the comparison ends
A small-scale mechanism.
The result applies to the tested seeds and porous models. A human parachute operates at a different scale and in a different aerodynamic regime, so this ratio cannot be transferred directly.
What the authors measured
The team combined free-fall tests, a vertical wind tunnel, particle image velocimetry and microfabricated porous disks. The ten biological samples supplied the means and confidence intervals shown here.
How to read the drawing
The animation shows the mechanism schematically through an original drawing. It does not reconstruct the authors’ velocity field, identify a detachment threshold or predict a seed’s trajectory.
Reproduce the area ratio
The reference disk is defined by the same weight, the same speed and the solid disk’s drag coefficient:
Ddisk = √[8mg ÷ (CDρπU²)] = 8.57 mm
Substituting the authors’ values in SI units gives a disk area of 57.67 mm² and a filament projected area of 12.56 mm²:
57.67 mm² ÷ 12.56 mm² = 4.59 > 4
The paper states “more than four times”. The values in the prose are rounded and should not be divided to claim extra precision.
Primary evidence
Where the data came from
- Cummins et al., A separated vortex ring underlies the flight of the dandelion, Nature 562 (2018) The primary paper, DOI 10.1038/s41586-018-0604-2.
- Full accepted manuscript Paper text, methods, extended figures and confidence intervals.
- The authors’ MATLAB scripts Edinburgh DataShare, CC BY 4.0, DOI 10.7488/ds/2362.
- The authors’ video data Videos of the 75% porous disk, Edinburgh DataShare, CC BY 4.0, DOI 10.7488/ds/2363.
Sources checked on 23 August 2026. The illustrations and calculations on this page are original; experimental values are attributed to the study’s authors.