The adaptive function of striping in the plains zebra

Wow! An incredible Post!

It was probably Parahyaena brunnea

Another very interesting read. Thank you. I am a complete amateur but have been interested in animal sight for some time; mammals, but also more recently, arthropods. Mammal colouration is inevitably discussed by guides in reserves and in popular literature in the context of human sight. I have always felt that it is far more relevant to try to understand it in the context of mammal (predator/prey) sight. Your post on the flags of lions and cheetah was also very interesting in this regard. I look forward to learning more.

@tonyrebelo

The main problem with the explanation invoking insects is that it assumes something special in the relationship between subSaharan equids and biting flies. What would be special here, accounting for a special adaptation seen neither in other ungulates in Africa, nor in equids elsewhere?

My explanation invokes something special in the relationship between the plains zebra and the spotted hyena. This is the combination of
1 a particular predator with certain characteristics different from those of other Carnivora,
2 a particular reliance, by the predator, on singling out a vulnerable individual for attack,
3 a particular scale in the images involved (a predator with a certain optical power scans prey of a certain body size, typically at a certain range of distances), and
4 a particular demographic liability on the part of the prey.

The plains zebra shimmers in the eyes of the spotted hyena because of the particular combination of the size and spacing of the stripes, the range at which the stripes are scanned, and the optical properties of the retina of the predator.

What the zebra is doing, in adaptive terms, is basically to compensate for its disadvantage in terms of an Africa-specific demographic liability, by means of an advantage in 'disarming' the scanning mechanism of the predator.

The latter is best understood in terms of the economics of time. The spotted hyena must not only scan before attacking, it must achieve its scanning at a certain rate. What the stripes hypothetically achieve is to retard this scanning, causing the predator to shift its scanning instead to adjacent prey species - i.e. ruminants - that do not have the demographic liability.

The 'achilles heel' of the plains zebra is its slow reproduction. The 'achilles heel' of the spotted hyena is its flicker-sensitive retina. By putting these together, striping allows these two species to continue to coexist.

When skunks display their striping, the context is different in terms of scale, range, and the optical systems of the potential predators. The striping of the skunk is too spaced-out, at the close range involved, for the viewer to perceive any shimmer. The skunk is not trying to hide subtle aspects of posture and movement, it is simply self-advertising (ominously) in a general way. The potential predators involved are various, with various optical adaptations. The nature of the message is different, in the same sense that the signal given by a 'zebra crossing' for pedestrians is different, in the eyes of a human driver, from the signal given by the optical puzzle I introduced at the start of this Post.

In summary, it is not that striping is confusing per se. It is only confusing in the special context of this particular interaction. And it is not the case that scrambling the retinal signal of the predator would be the best protection for other species of prey, or for prey generally. It is the best tactic in this particular case because of the particular context of risk/reward, or cost/benefit.

The plains zebra would not be especially liable demographically (to inadvertent victimisation) were it not for its inferior reproductive rate relative to the predominant species of prey in its ecosystem. The spotted hyena would not be especially liable optically (to 'retinal scrambling') were it not for its particular optical system and reliance on targeting the unfit. The evolutionary interplay between these two liabilities, on the part of prey and predator, provides a robust conceptual framework that I have not found in any other explanation of the striping of the plains zebra, including the one invoking biting flies.

Yet another aspect is as follows: A basic pattern in 'plains ruminants' is that individuals stot when scanning predators arrive. The tactic is to shift the attention of the predator away from the obviously-fit and on to the less fit. Stotting cooperates with the visual system of predators to ensure that the least-fit are targeted, which makes demographic sense and serves both predator and prey.

The tactic of the plains zebra also involves a shifting of attention, but in a sense by an opposite method to that seen in stotting (which the plains zebra never performs, even as an exuberant juvenile). Instead of enhancing the visual system of the predator in discerning individual fitness, the plains zebra hinders it. The effect is similar in that the attention of the predator is shifted. But the crucial difference is that, in this case, it tends to be shifted not from a fit individual ruminant to an unfit one, but instead from the plains zebra as a category to ruminants as a category. This makes sense because, in a sense, the whole population of the plains zebra is 'unfit' to start with, owing to its inability to sustain the offtake rates normal for the ruminants around it.

Does this help to explain?

You merely reiterated above what you said earlier, an argument that I accepted. But the Tsetse Fly hypothesis is more parsimonious, requiring fewer assumptions (1. zebras are particularly susceptible to Trypanosoma, and 2. shimmering out of Tsetse vision is a solution), and is tested (fly traps painted with stripes catch less flies), and offers predictions: no Tsetse - no stripes: works with Equus quagga sensu latu, but E. zebra is a problem. Zebras are not listed as Tsetse hosts (https://en.wikipedia.org/wiki/Tsetse_fly)
However, Spotted Hyenas occured to Cape Town, where Zebras do not shimmer, and they extended to Asia, where no equids went stripy/shimmery.
Einsteins Razor: any theory must be as simple as possible and no simpler. I think the balance favours Tsetse over Hyenas, although I still think - if one solution solves two problems, then go for it.

@tonyrebelo

Why, at a basic ecological/physiological level, would the plains zebra be particularly susceptible to Trypanosoma? And, if it was, would the evolutionarily parsimonious adaptive option not be simply to improve immunity at the level of the physiological immune system?

Is it really scientifically parsimonious to invoke a parasitic insect, or the parasitic disease it transmits, as constituting a selective pressure so great that it shapes a major aspect of the appearance of a large-bodied animal? Flies and trypanosomes certainly reduce the vigour of their victims, but can this really be considered on a par with a major predator, which kills outright?

We see, everywhere in the animal world, major adaptations to predation, from the anatomy of limbs to the colouration of skunks, and from the armour of armadillos to the stotting of gazelles. Subordinate only to nutrition and sex, the animal world seems to revolve around predation - a powerful-enough phenomenon to shape animals profoundly.

Is there any comparable adaptive power in the phenomenon of parasitism? What would be an example of a mammalian species the basic appearance of which is shaped by anti-parasite adaptation?

A problem with invoking flies in explanation of the striping of the plains zebra is that this striping has major effects on various other adaptive options in the life of the equid. To the degree that it affects predation, surely the costs might exceed the benefits?

I.e. there is surely a hierarchy of importance, and a set of priorities, in overall adaptation? First come the lethal risks, then come the minor risks. Is invoking flies not somewhat akin to a 'tail wags the dog' rationale?

The extinct quagga is indeed puzzling w.r.t. the relationship of the plains zebra to the spotted hyena. This does deserve more thought.

However, my biggest problem with Caro's idea w.r.t. flies is that I know of no other case of a major adaptive feature in a large-bodied vertebrate being profoundly modified by selective pressure from a microorganism. In that sense, Caro's hypothesis seems fundamentally non-parsimonious to me. Flies do not seem to have much explanatory value, because no basic energetic/evolutionary principles seem to be revealed. And this problem is hardly mitigated by his experimental tests, because in Science the conceptual framework is more important than experimental verification. No experiment means much except with reference to a particular set of principles.

Your further thoughts?

@beartracker Wildebeests (Connochaetes) and the plains zebra live together, face the same predators, and have extremely precocial newborns. However, they differ in stotting behaviour. The plains zebra does not stot even in infancy, whereas the following shows that Connochaetes mearnsi does so: https://www.youtube.com/watch?v=wq-ScQkDgNM. The following shows that play-running in the plains zebra is exuberant, but does not involve any special gait such as bouncing or style-trotting: https://www.youtube.com/watch?v=J4flNLIxsPA. Here is comparable footage for Grevy's zebra: https://www.youtube.com/watch?v=RVgaCKeJQmI and https://www.youtube.com/watch?v=ncMW41SLAl0v.

THINKING THROUGH THE TOPIC OF DEMOGRAPHIC LIABILITY

Equids tend to reproduce more slowly than like-size ruminants which outnumber them in their African habitats. Because the carrying capacity for predators depends mainly on the ruminants (which is simply because the ruminants provide most of the food for the predators), the equids will tend eventually to be exterminated if they are predated in proportion to their numbers.
 
Let us say the prey community consists of just one species of wildebeest (80% of total population) and one species of zebra (20% of total population). The reproductive = offtake rate of the whole prey community, which maintains constant density from year to year, is x. This value, x, is determined mainly by the wildebeest in our example community. Equids have a reproductive rate of y, a value smaller than x. If predators take the zebra at a rate of one in five kills (in proportion to their contribution to the instantaneous population), then the zebra will be depleted at the rate of x-y even as the wildebeest slowly and slightly increases its numbers.

Eventually the proportion of the population consisting of the zebra will be 0%. The predation has been fully sustainable for the prey community as a whole, but the equid has disappeared from the community.
 
The idea is that, in a mathematical sense, the zebra is being discriminated against. The predators are not particularly victimising the zebra but the zebra cannot replace itself at a sustainable rate even when it is picked off at random from a collective population on a level playing field. In order to survive, it has to put more into protecting and defending itself from predators than the wildebeest does. Since the equid cannot boost its reproductive rate, if it is to survive it must somehow lower its predation rate.
 
The zebra cannot outrun the ruminants, which are already world-beaters in speed and endurance. However, it can be more extreme than the wildebeest in several features relevant to predation, such as:

 a) precociality of neonates (wildebeests too have remarkably precocial neonates, but the size relative to maternal size is greater in zebras) 

b) defence of infants by the whole group, not just the mother (as far as I know no wildebeest emulates zebras in this way) 

c) kicking (more powerful and directed than in like-size ruminants, particularly wildebeests) 

d) skin-shield (a minor form of armour) on the rump (although wildebeests are extremely tough animals, e.g. when shot with bullets, as far as I know there is nothing special about the skin on the rump) 

e) defence of females by males despite sexual monomorphism (as far as I know, wildebeests show no defence of others by males, even though some forms of wildebeest are among the feistier of antelopes vs predators by means of their horns) 

f) vigilance (zebras have proportionately larger eyes than those of wildebeests)

g) refuge behaviour at night (as far as I know, wildebeests do not resort to bedding areas with particularly short, open vegetation to the degree that some zebras do) 

h) alarm-vocalisation (although wildebeests snort in alarm, they do not vocalise in alarm and even their social vocalisations are relatively soft) 

i) seeking proximity to wildebeests and other ruminants (it seems to be zebras that seek out ruminants, rather than vice-versa).  

It is to this list that I add the striping, which can be thought of as ‘the visual equivalent of armour’ in the sense that it tends to make the zebra ‘impenetrable’ to the gaze of certain predators. 

(continued in next comment...)

Here is unusually clear footage of the Asian wild ass: https://www.youtube.com/watch?v=2Gydq5nSSYk.

I've seen that alarm walking behavior many times in blacktailed deer. They are the only deer species in my area. They hold their head up and walk with a sort of stiff-legged gait. It's interesting to watch.

The dont need the tail as the flies cannot see them ....

Interestingly, bobcats have black backs of their ears, with a white spot in the center. Maybe because the markings look like eyes?

@beartracker Also https://www.inaturalist.org/journal/milewski/54939-felid-eyespots-are-like-human-eyes-but-not-in-the-way-you-think# and https://www.inaturalist.org/journal/milewski/54980-if-felid-eyespots-lead-offspring-why-are-they-already-present-at-birth#

The flicker-effect of zebra striping in the eyes of Carnivora is crucial for my explanation. Does our knowledge of vision in the domestic dog support the idea that Carnivora tend to see flickering even where the human eye would not?
 
The summary below suggests that the answer is yes. In addition to being more motion-sensitive and light-sensitive than the human eye, the dog’s eye is also known to be more flicker-sensitive specifically.
 
This is support for the idea that the spotted hyena is likely to see flickering when observing a stationary zebra, provided that one is prepared to extrapolate from canids to hyenids.
 
What the account below suggests is that when human and dog sit watching TV, the experience is pleasant and informative for the human, but not for the dog. The reason is that what the dog sees is not moving pictures as much as an interference effect.
 
True, a difference between TV screen and stationary zebra is that the former has moving stimuli, whereas the latter has stationary stimuli. However, the crucial fact here is that even the human eye sometimes sees an ‘illusory’ flicker when viewing zebra striping, as I have shown with both still pictures and video footage. Given that even our eye sees flickering in what is actually a stationary scene, it seems likely – putting these two lines of evidence together – that the carnivore will see a disconcerting degree of flickering when watching a stationary or slow-moving zebra.
 
http://www.dogwalkersmelbourne.com.au/articles-dog-walking-pet-sitting/66-dog-canine-vision-seeing-compare-human

Explanation of how the domestic dog sees striping:
 
Please see http://www.eyevet.ca/coile1.html and https://www.dogwalkersmelbourne.com.au/articles-dog-walking-pet-sitting/66-dog-canine-vision-seeing-compare-human.
 
The idea is to imagine a full moon, as seen in the sky by the human eye or the dog’s eye.
 
The human eye, when emmetropic (https://en.wikipedia.org/wiki/Emmetropia), can discern about 150 stripes packed into the area of the full moon. (It would be interesting to calculate the distance to which this corresponds in the case of a human spotting a distant zebra, but I do not know the relevant factor.)
 
Now, the domestic dog can only discern about 30 stripes on the same area of a full moon. It seems to be unknown to what degree this limitation is the result of simply inferior visual acuity, vs a flicker-effect that overrides visual acuity. (The reference below, although sound, does not invoke any flicker-effect in this comparison.)
 
Although it may seem competent that the dog can look at an area as small as the full moon in the sky and discern a hypothetical 30 stripes on this area, its ability is about 5-fold inferior to the human ability in this regard.
 
So, the human eye will be able to discern striping on a distant zebra much farther away than is possible for a dog – despite the fact that the pupil of dog is larger than that of human.
 
The spotted hyena has absolutely larger eyes than that of the dog, but its value is similar: about 30 stripes on the full moon. This is because the spotted hyena - although it often hunts and perform other activities by day - has a retina dominated by rods (https://www.cis.rit.edu/people/faculty/montag/vandplite/pages/chap_9/ch9p1.html#:~:text=Rods%20are%20responsible%20for%20vision,responsible%20for%20high%20spatial%20acuity.), reducing its visual acuity in a way typical for nocturnal mammals (Calderone et al. 2003).

Although it seems likely that Carnivora can see zebra striping up-close (e.g. a few tens of metres distant), it remains unknown at which distance the hypothesised flicker-effect sets in. At some range of distances it seems reasonable to assume that the dog will see the zebra as striped, but with the stripes flickering.

@tonyrebelo @jacqueline_llerena @beartracker

Visual acuity (https://assessmentofequinebehaviour.wordpress.com/perception/visual-acuity/) in the cursorial carnivores that prey on zebras:
  
A human with normal visual acuity has 20/20 vision (able to discern lines of standard separation at a distance of 20 feet). The domestic dog has visual acuity of only about 20/80.

Whereas an average human (with normal healthy eyes, and in normal daylight) can distinguish lines at a range of 80 feet, the dog cannot do so beyond 20 feet. Because the visual acuity of dog is only about a quarter of the human value, the dog needs to be much closer to the lines to see that they are lines in the first place (as opposed to a grey blur).

(Some hawks have visual acuity of 20/2, which means that they can see the same lines from an order of magnitude farther away than is possible for a normal average human with healthy eyes.)
 
The wolf may exceed the domestic dog in visual acuity. However, it probably remains far inferior to the human in normal daylight, in this particular parameter. In dim light, the visual acuity of dog and wolf are of course superior to ours. However, lines are so hard to tell apart in dim light that dog/wolf probably still cannot distinguish lines at any distance to speak of.
 
This means that objects that humans can see to be striped, in normal daylight, would tend to lose their stripiness in the wolf’s eyes at the same distance. And the ability of the wolf to see well in dim light would not much alter this.
 
At any distance, the stripes of zebras are probably invisible to the wolf. How far does a zebra have to be from a normal human with healthy eyes, before its stripes blur into a greyish smudge. 100 m? If so, then a wolf probably loses sight of the striping at distances > 50 m, or perhaps even > 30m?
 
This may help to explain how zebras contrive to use a disruptive pattern (in our eyes) to achieve a cryptic pattern (in the wolf's eyes). To their natural predators, zebras possibly look stripeless and greyish at any distance - except for those parts of the body on which the striping is particularly bold. When they approach closely enough to see the striping, the disconcerting flicker effect presumably arises.
 
Felids are inferior again to canids, in visual acuity. Thus, a lion probably needs to be even closer to a zebra than does an African hunting dog, in order to discern stripiness.
 
The visual acuity of the domestic horse is said to be 20/30, superior to canids and felids, but inferior to humans. Where a normal human with healthy eyes can tell lines apart at 30 feet, the domestic horse needs to be 20 feet from the lines to see them as lines, as opposed to a greyish blur. So, zebras are far more likely to see each other’s stripes, at any distance, than their canid predators are, and far more likely than their felid predators are.
 
The horse has visual acuity inferior to that of humans, despite a much larger eye, and partly because it lacks a fovea (https://en.wikipedia.org/wiki/Fovea_centralis). The horse instead has a retina with the most sensitive part arranged in a horizontal ‘streak’. The horizontal alignment allows the horse to monitor most of the horizon around the animal without moving the head. The horse, in a way, trades quality of vision for quantity of vision. It sees over a wider horizon than we do, but with less acuity.

Confirming that vision in felids is similar to that in hyenids and canids: 
 
Of all the books I possess, perhaps the best in describing and explaining vision in felids is ‘Cat sense: inside the feline mind’ (1994) by Pirincci A & Degen R, Fourth Estate, London.
 
The cheetah has forward-facing eyes in a human-like face, and seems to focus intently on its prey with circular pupils (unlike the vertical slit pupils of smaller felids). However, even this species lacks the fovea so important in the human retina. Instead, like the wolf, the retina of the cheetah has a horizontal streak (more like that of its prey the gazelles). This enables it to scan a horizon for movement, rather than to study any specific object by focussing on it. So a cheetah can possibly look right at a zebra and not see it, unless it moves. This is true despite the fact that the cheetah has excellent vision, and depth-perception.
 
Felids assess distances not purely by our binocular method, but partly by moving their heads in a small arc. This is well-known in the domestic cat. Felids are good at judging distance when pouncing on prey or jumping from branch to branch. However, this does not necessarily mean that they can focus on an object, in the way we do.
 
Although felids are more binocular than canids (and I think the striped hyena is more binocular than the spotted hyena, judging from photos), it is nevertheless true that felids retain more peripheral vision that humans do. This seems to apply even to the largest felids, such as the lion. Felids simply cannot focus on, and study, stationary objects in the way that humans take for granted. The felid eye remains, like that of its prey and other Carnivora, mainly designed to detect movement and to do so even in dim light.
 
All Carnivora, then, have eyes different from human eyes, in being far more sensitive to movement at the expense of an ability to study stationary objects. There is variation among Carnivora, with some forms more binocular than others. However, binocularity, as seen in felids, is limited. It can be misleading, if one assumes that it means much convergence with the foveal focus of vision so typical of humans. Even to the lion or the cheetah, the zebra is seen quite differently from the way we see it. My readings indicate that any differences between spotted hyena and lion are minor, and that the same explanation for the adaptive value of striping in zebras would apply to all the non-human predators of zebras.

I like the idea behind http://www.businessinsider.com.au/pictures-of-how-cats-see-the-world-2013-10. They have tried to show, in the form of actual picture comparisons, how felids see the world relative to how humans see the same scenes.

I could quibble with some details (e.g. I do not think that the night-vision pictures adequately show that felids see much more in near-darkness than we do). However, one still gets a good idea of how BLURRY felid vision is, compared to human vision - under almost all circumstances. Felids do not care about this blurriness, because they are not trying to focus on objects as we do. The lack of visual acuity and focus is more than compensated, in a felid’s sensory world, by a) extremely good detection of rapid movements (far superior to ours), b) extremely good detection of movement even in the dark (ever farther superior to ours), and c) a sense of hearing far superior to ours.
 
Surprising as it may seem, a domestic cat cannot focus, either on its human’s face, or on objects a hundred metres distant.
 
This means that the striping of zebras will surely look blurred to a felid, either in a static way or with flickering.
 
The flickering can be regarded as an ‘anti-focus’ effect, i.e. an actual aggravation of the already poor ability of the felid eye to focus. Without getting too technical, the reason why felids see as flickering things that we see as steady (e.g. a TV screen) is that their retinas are so extremely sensitive to movement that they detect motion that our eyes overlook, in space as well as in time.

Also see https://www.inaturalist.org/journal/milewski/68153-significant-that-when-spotted-hyena-hunts-plains-zebra-it-is-males-that-tend-to-do-the-hunting#.

Lateral inhibition! Weird!

Dont forget that this was the reasoning for Tsetse Flies not being able to detect Zebras!

https://theconversation.com/zebras-stripes-are-a-no-fly-zone-for-flies-111888
"Early experiments in the 1980s ... tsetse flies and horseflies avoid landing on striped surfaces and .... confirmed more recently."
"... across the geographic range of the seven living species of equids. ... intensity of striping closely parallels biting fly annoyance in Africa and Asia ... annoyance from horseflies is prolonged over the year ... likely to have marked striping patterns."

https://resjournals.onlinelibrary.wiley.com/doi/10.1111/j.1365-3032.1992.tb01191.x
two species of tsetse ... the vertically-striped target was significantly less attractive than black or white (P < 0.001).

Yes, predators may kill x% of the population per year, but biting flies, potentially transmitting diseases will bite 100% of the population many times per season.
In the overall cost/benefit regime of adaptive options, carnivores as outright killers of a few animals, are likely to be subordinate to massive inoculation by hoards of biting flies.

There is no point arguing these issues. Just design an experiment to exclusively test between these two hypotheses. Given your optical explanations, it should be possible to easily test this using videos on lions, hyenas and flies in zoos.

Photo illustrating the topic of predation by spotted hyena on plains zebra vs wildebeests:
 
This following is particularly nice in showing how the plains zebra looks inconspicuous next to the conspicuously dark wildebeest.
  
Crocuta crocuta, Connochaetes mearnsi and Equus quagga boehmi:
http://www.travelagewest.com/uploadedImages/TAW_Photo_Tours_-_Videos/Africa-Middle_East/750x550_150316_CS_PhotoTour_7.jpg

When the spotted hyena surveys this scene in search of a vulnerable individual to chase, its eyes may tend to be drawn to the wildebeest. This is for three reasons, viz.: a) wildebeest > zebra in numbers, b) wildebeest more cinspicuous, thus easier to spot, than zebra, and c) zebra harder to assess for individual vulnerability owing to confusing flicker of striping in the eyes of hyena.

Please see https://explorebioedge.com/2020/06/15/__trashed-32/.