Every claim in the Short, and the sentence behind it
C1: Horseflies still came close
Horseflies circled zebras and horses at similar rates. Stripes did not stop them approaching.
no significant difference in the rates at which tabanids circled zebras or horses
Source: Results.
C2: Far fewer flies landed on zebras
The difference showed up when the flies tried to land.
significantly fewer tabanids landed on zebras than on horses
Source: Results.
C3: Flies failed to slow down before reaching a zebra
Flies approaching zebras did not slow down in a controlled way at the end of their flight.
tabanids approaching zebras failed to decelerate in a controlled fashion towards the end of their flight trajectories
Source: Results.
C4: A smaller share managed to land
The reported landing proportions were about 9.9% on zebras and 27.9% on horses. Flies often bumped into the zebra instead.
a significantly lower proportion landed on zebras than on horses
Source: Results.
C5: Striped coats helped horses too
Horses wearing striped coats got far fewer touches and landings on the coats than horses wearing plain black or white coats.
with far fewer touching and landing on striped than uniform coats
Source: Results.
C6: The bare head still got landings
The horses' uncovered heads got landed on just as often whatever coat they wore. The result points to an effect at close range.
the zebra cloth coat had beneficial effects for the horse but the naked head suffered the same frequency of landings by tabanids
Source: Results.
C7: No fly was seen probing zebra skin
Observers saw no horsefly probe zebra skin during 5.3 hours of direct observation. They saw 239 instances on horses during 11.0 hours.
we saw no tabanid probe zebra skin during 5.3 hours total of direct zebra observation as opposed to 239 instances during 11.0 hours of observing horses
Source: Results.
C8: Filmed zebra landings were brief
Flies stayed an average of 1.20 seconds on zebras and 10.06 seconds on horses. The zebra average rests on just five filmed landings; the horse average on 886.
tabanids spent 1.20 sec (SE = 0.91, N = 5 instances) on zebras but an average of 10.06 sec (SE = 0.08, N = 886 instances) on horses
Source: Results.
C9: The authors suggest stripes become visible up close
The authors suggest horseflies may only make out the stripes from about two metres away.
Our data suggest that tabanids may only be able to resolve stripes at a 2m distance
Source: Discussion.
C10: This happened on an English horse farm
The study watched captive zebras at Hill Livery in Dundry, North Somerset, UK.
at the Hill Livery, Dundry, North Somerset, UK
Source: Methods.
What the study doesn't say
"Forget how to land" describes the failed landing. The study does not establish that a fly forgets anything or explain why stripes interfere with landing.
The direct observations involved three captive female plains zebras and nine horses. Seven horses took part in the coat experiment. Only five zebra landings were caught on video, so the comparison of time spent on the animals has a very small zebra sample.
These were European horseflies on a farm in Somerset. African flies were not tested. Observations took place in calm, dry weather.
Zebras also flicked their tails more and walked or ran away from flies. The absence of observed probing cannot be credited to stripes alone.
The comparison had other limits. A single grey horse had lower circling and touching rates than the zebras. The fields differed in size, and the coats were different fabrics bought off the shelf. The coat observers knew the hypotheses; most video coders did not.
The paper's phrase "almost one third fewer" does not accurately describe its reported landing proportions. The figures above are quoted as percentages instead.
The study does not prove that biting flies are the evolutionary reason zebras have stripes. A later review says fly deterrence has the strongest empirical support, but cannot alone explain why stripes are strong in zebras and not in other African mammals.
Later experiments also weaken the original explanation for the failed slowdown. Horseflies avoided checked patterns too, which goes against the proposed aperture effect. Another study found avoidance of high-contrast stripes, but not low-contrast ones; its working hypothesis concerns the absence of large dark patches that attract flies. A later field study found close-range repulsion across zebra species. These follow-up accounts are based on the studies' abstracts, not on checks of the full papers.
The source
Caro T, Argueta Y, Briolat ES, Bruggink J, Kasprowsky M, Lake J, Mitchell MJ, Richardson S, How M. "Benefits of zebra stripes: Behaviour of tabanid flies around zebras and horses." PLOS ONE, 20 February 2019.
DOI: 10.1371/journal.pone.0210831.
Read the paper. Full-text XML.
Licence: CC BY 4.0. This page quotes and links to the paper. It reuses no figures.
Later research on the same question:
- How et al., 2020: checked patterns and the aperture effect.
- Caro et al., 2023: stripe contrast and dark patches.
- Tombak et al., 2022: close-range repulsion across zebra species.
- Ireland and Ruxton, 2025: review of stripe explanations.
The Short
Watch it on YouTube: https://youtu.be/RwkUQh4kiU8