Thursday, 25 July 2013

The Quagga and the Quagga Project

The Quagga, Equus quagga quagga, got some attention during the last decade because of a selective breeding project in South Africa that tries to “rebreed” the appearance of this Plains zebra subspecies. The Quagga project is different from those concerning the aurochs or wild horse, particularly because there are no living descendants of the Quagga to work on. The identity and evolutional history of the Quagga has been a matter of debate for a considerable time, but let’s have a look at the history of the equines referred to as zebras as a whole, because we have to look at the complete picture in order to understand it better.

Surprisingly, a study by Orlando et al. 2009 [1] found zebras to be a polyphyletic arrangement, with the Mountain Zebra E. zebra being more closely related to the African ass E. asinus, the Grevy’s zebra E. grevi within hemionine asses and the Plains zebra as an outgroup to these two clades. According to this study, zebras are an umbrella term for heavily striped wild asses, if you will. But the paper itself states these conclusions are to be taken with caution, and mtDNA analyses might be necessary for better resolving the non-caballine Equus tree. Another 2009 study finds zebras to be monophyletic, but oddly the onager outside E. hemionus [2].
The phylogeny of zebras is essential for understanding the evolutional cause of their striped coats, in order to deduce which coat pattern is a plesiomorphy and which is an apomorphy. If the plesiomorphic state of the clade formed by zebras and their most recent common ancestor is a bay dun coat colour, it means that the striped zebra coat pattern evolved several times independently within Africa (E. simplicidens from the Pleistocene of America is often claimed to be a zebra-like animal, but I think this is highly speculative based on skeletal morphology alone). If the plesiomorphic state is the striped coat, it must have been lost several times again and re-developed into a bay dun coat of the asses nested within that group according to Orlando et al. 2009. Several hypotheses try to explain why zebras developed their unique coat, and if it evolved several times independently within Africa, a connection to a special environmental condition is very likely. The hypothesis that it evolved in order to remain undetected for tsetse flies and horseflies is the most popular one, and indeed it has been shown that polarized striped patterns is less attractive to tabanid flies [3]. According to the hypothesis, zebras needed this protection because their relatively recent arrival on the African continent did not give them the time to evolve efficient immune defence against diseases transmitted by these parasites. And that is how we finally come to the Quagga: far in the south, where the tsetse fly is not present, the stripe pattern allegedly got useless and therefore these zebras reduced it again. But I have several problems with this scenario. First of all, there is no real correlation between the range of the tsetse flies and that of the zebras. The modern range of the Mountain zebra’s range is totally outside (and it actually occurs as southwards as the Quagga did), as is a large part of that of the Plains zebra, and I wouldn’t be surprised if the former range of the African wild ass and that of the tsetse fly also overlapped. Furthermore, I don’t know of studies suggesting that zebras are more vulnerable to trypanosomiasis than other African herbivores, making the passive protection through their striped coat necessary. I think the “fly hypothesis” is plausible, but not yet really convincing for me (please correct me if my criticism is erroneous).

Regardless of why the Quagga had reduced stripes, they were its most remarkable feature (there is the possibility that this condition actually is a basal one, but I’ll address that later on). Sometimes the quagga is described as looking like half zebra and half horse, but I don’t think that is an ideal portrayal of its appearance. I would rather describe it as a brown zebra with white legs and tail and stripes fading away at the posterior half of its body. On its brownish rear, the Quagga displayed an eel stripe just like phenotypically bay dun wild equines. Most stuffed specimen show brownish stripes, but that’s likely due to discolouration because all of them are more than hundred years old. Based on contemporaneous paintings and photographs of one individual, I think that the lighter space between the stripes on the neck and head area were white or whitish, in some individuals maybe brownish. With this (baaadly rendered) drawing, which is also featured on the Quagga Project’s website, I tried to portray the individual variation within the Quagga:

The preserved specimen of the Quagga, which are all listed here, might give you a good idea of the variation within the population as well. Considering the unique pelage characteristics and also alleged craniometric differences between the Quagga and the Plains zebra [4]*, it is understandable that it has been regarded as a separate species for a long time. However, genetic studies found the Quagga to cluster within the Plains zebra (thus indicating that its ancestors were fully striped) and having little genetic diversity. It is likely that the Quagga split from other lineages between 290.000 and 120.000 years ago, possibly due reproductive isolation during the last glacial, what also provides a possible scenario for the radically different pelage colour (genetic drift) [3]. Most importantly, the Quagga does not share any haplotypes with other Plains zebras according to this study, indicating that there was no genetic exchange after their isolation [4]. This means that the Quagga is a genetically distinct subspecies rather than a mere colour variant, although more material might be needed.

* Some have disputed the craniometric differences because available horse or donkey skulls were used for some mounted zebra skins instead of genuine skulls. However, I know that there is at least one preserved Quagga skull (on the only mounted Quagga skeleton); I wonder which material those cranimetric comparisons are based on. If you want to judge it for yourself, here is the confirmed Quagga skeleton with the skull in profile view, and a Plains zebra skull.

However, the Quagga’s outer appearance was not homogeneous, there actually was considerable variation concerning the degree of striping among the mounted Quagga skins. Some museum specimen show a very brownish coat with stripes only recognizable at the shoulder, neck and head area, others display stripes that go right to the hips and a comparably light background colouration (as far as we can tell from those skins). Interestingly, the living plains zebras also show a degree of striping varying from north to south, with those in the north having the most polarized stripe pattern (E. q. boehmi) and those in the south showing a tendency to reduced striping on the rear and legs (E. q. burchelli). Concerning the pelage characteristics, the quagga might represent the end of a cline, although it seems to be genetically distinct.

Because of the obvious variation in modern Burchell’s Zebras striping, Lutz Heck proposed that selective breeding could rebreed the Quagga’s outer appearance (phenotype) in a book published in 1955. Eventually a project with this aim was formed by zoologists, veterinarians and museum personnel under the initiative of Reinhold Rau in 1987 [4]. They selected wild zebras, mostly from the Etosha national park and exclusively from the burchelli subspecies that already showed a promising-looking amount of stripe-reduction. The current herd has 83 of zebras resulting from this selective breeding programme living in several locations near Cape Town. The youngest individuals are from the fourth generation. Keep in mind that zebras have a longer generation span than domestic horses so breeding with them takes longer. The project selected for a reduced striping and a more brownish background colour by crossing the chosen wild zebras and choosing the offspring with the least striping for further breeding. Surprisingly, some animals also showed stronger striping, but those were selected out. The breeding progressed surprisingly fast, already the third generation showd considerable reduction of stripes on the legs and rear, a white tail and a slightly more brownish background colour. The fourth generation (some of them were born in 2012) show an even greater approximation towards the Quagga’s pelage characteristics. The project developed a counting system in order to quantify the progress, and the results show that the reduction of striping is happening fast, but the darkening of the background colour progresses only slowly (this aspect is especially hard to quantify because of dust, sun et cetera). Here are some photos and a video of some of these “Rau zebras”**:

Individual "Freddy", F4
Individual "Henry", F3
Individual FM12, F4

** Funnily, I independently came up with the idea of calling these animals “Rau zebra”, the Quagga Project itself wants to name their results “Rau Quagga”, but I think calling them zebras is more objective.

If you are interesting in seeing more photos of these Rau Zebras, you can have a look at the homepage or the project’s facebook page. In contrast to aurochs or wild horse projects, where all desired phenotypic features are present in living animals (which descend from the desired archetype) and “just” have the be united, the Quagga Project wants to achieve features that are not present in any living zebras (which do not descend from the Quagga). If you’d imagine a bell curve, the Boehm zebra is on the one end and the Quagga is on the other end, and the Burchell zebras are somewhere in between. Nevertheless, the optical resemblance that already has been achieved is impressive to me. Some of these individuals have show a significant resemblance to the Quagga already, like “Freddy”, “Henry”, DJ10 and FM12. The stripes on the legs are greatly reduced or absent, the stripes on the rear are reduced, and the background colour is more brownish than on average Bruchells zebras. Imagine how Quagga-like the F5 and F6 generation might become. But can the project truly rebreed all the Quagga’s phenotypic features, or even the Quagga as a whole?
The Quagga’s appearance often is equalized with its coat, but there are other features that yet have to be evaluated. For example, there is no osteometric study working with that one complete skeleton and that of living Plains zebras, and since the use of skulls from stuffed skins is questionable, one should compare the genuine Quagga skull with that of living zebras in order to verify the statement that there are no osteometric/craniometric differences between the Quagga and other members of that species. I think it is well plausible that there are, because of the apparent founder effect during its evolution, but this simply requires testing.
Furthermore, there is more that defines a subspecies than just a differing phenotype. The argument “the Quagga was merely a subspecies, so it can be rebred” that is sometimes brought up in connection with the project is erroneous. Some of the Youtube commenters seemingly have a very simplified idea of species and little clue about population genetics. Species that are divided into several subspecies are not genetically homogenous, connected populations but more like a branched tree, with each subspecies being an evolutionary distinct clade, defined by whatever differences that are laid down in the genome. And apparently there was not much genetic exchange between the Quagga and other Plains zebra populations for at least 120.000 years (although more evidence might be needed). This means that E. q. quagga was genetically different from E. q. burchelli just like burchelli is from E. boehmi and so on. To make a comparison on a similar level, you cannot rebreed the Polar wolf by selecting white Timber wolves. Also, you cannot rebreed the Boehm Zebra by selecting those Burchell’s Zebra with the heaviest striping. Different subspecies mean a different genetic make-up (although not as marked as on species level, of course) and therefore a different animal. Consequently, the Rau zebra will always differ from the Quagga because the populations have a different ancestral history. These differences are at least genetically, and perhaps also in ecologic, behavioural or phenotypic respects that are either unknown or not evaluated, although I personally think these aspects are overall roughly the same the within the Plains zebra.
We probably cannot say anything about the exact ecologic role of the Quagga because there are no living representatives left to study. But probably, since they belong to the same species, they were largely similar if not ecologically identical to other Plains zebras. I read in some sources that South Africa’s flora has a higher degree of endemism than farther in the north, but I don’t know if that is true (does anybody have literature on that?). The Quagga project claims that the flora is not significantly different from that of the habitat of living members of the species. I think that many conservationist would be in favour of reintroducing the Plains zebra into its former range in South Africa, as subspecies are obviously no barrier for reintroduction projects, especially when the native type is lost (as the reintroduction of Przewalski horses in some parts of Europe or recent projects with big cats have shown). The ecologic functionality and physical characteristics of the Rau zebras are the same as in other plains zebras, but they are the only lineage that resembles the type that once was native in that region. If the Plains zebra is to return into this area – what would surely make sense from a conservational view – the Rau zebras from the Quagga Project certainly are the most desirable option (just like the project says on their web page).
I think it is very helpful that the Quagga Project methodologically quantifies their selective breeding progress; also, it’s the only breeding-back project that has published papers explaining their “materials and methods”, making the whole process much more scientific and transparent (see here and here). Apart from that, I am also curious to see other effects on these zebras because of the selective breeding, and perhaps also new mutations will show up. 

Although the Quagga itself is lost, I am very happy that the Quagga Project is trying to establish zebras resembling this remarkable subspecies. It will be interesting to see how far they can get and indeed the similarities that already have been achieved are astonishing. Some of these Rau zebras already show an amount of stripe reduction that is shared by some Quagga mounts, and if they had the same brownish background colour they would be virtually indistinguishable. Because the progress in stripe reduction has been fast but not the change of the background colour, the project decided to put more focus on that feature [5].
Also, the Rau zebra would be a very useful tool if it will be possible to clone one or more Quagga individuals one day, however unlikely that scenario is. But these few cloned Quaggas could constantly be merged into the Rau zebra’s population that provides a genetic base, coming as close to the Quagga as technically possible. However, this is a fantasy scenario of mine (yet?).


  • [1] Orlando et al.: Revising the recent evolutionary history of equids using ancient DNA. 2009
  • [2] Samantha A. Price und Olaf R. P. Bininda-Emonds: A comprehensive phylogeny of extant horses, rhinos and tapirs (Perissodactyla) through data combination. 2009.
  • [2] Knight, Kathryn: How the Zebra Got its stripes. 2012.
  • [3] Leonard et al.: A rapid loss of stripes: the evolutionary history of the extinct quagga. 2005
  • [4] Eric Harley et al.: The restoration of the Quagga: 24 years of selective breeding. 2013.


  1. Are the aurochs back breeding programs using objective measurements to score individuals and to guide future breeding? If so, I would love to learn about that.

    1. Until now, no aurochs effigy-breeding projects do that; but it is possible that Tauros Project is going to do something like that. I already asked Henri Kerkdijk-Otten about that. IMO, Tauros Project yet needs to make their "material and methods" publicly, just like the Quagga Project did.