Sunday, 15 September 2019

Unravelling the Bushbuck

Three years ago, I wrote a series of posts on the various species of bovine. This was using 'bovine' in the wider, technical, sense than it's everyday meaning, essentially including all animals more closely related to cows than to sheep or goats. This obviously includes all the really close relatives of cattle - bison, buffalo, yak, and so on - but it also includes a number of animals that are more accurately described as 'antelopes'.

Most species of these 'bovine antelopes' belong to the genus Tragelaphus, collectively known as 'spiral-horned antelopes'. Especially if we include the large cow-like elands (sometimes given their own genus, Taurotragus) these are a fairly diverse range of animals, including some adapted to open woodland or savannah, and others adapted to dry scrub, mountain slopes, dense jungle, or swampland. Perhaps because it seems to be the most adaptable of the species, the most widespread of the spiral-horned antelopes is the bushbuck (Tragelaphus scriptus).

When I last wrote about the bushbuck, I concluded by mentioning the results of a 2008 study that had found something strange in their genetics. Specifically, it turned out that bushbucks could be divided into two distinct populations, one of which proved to be closely related to another spiral-horned species called the nyala, while the other was more closely related to the bongo and sitatunga. The only reasonable conclusion was that bushbucks, as we currently understand them, are actually two species, not one.

The "new" one, which lives further south and to the east than the animal originally given the name in the 18th century, has been referred to as the imbabala or Cape bushbuck (Tragelaphus sylvaticus).

In some ways, this wasn't hugely surprising, because there's plenty of precedent for it. We already knew, for example, that despite their similar appearance, the nyala (T. angasi) and mountain nyala (T. buxtoni) are not each other's closest relatives. Another pair of very similar-looking spiral-horned antelopes, the greater kudu (T. strepsiceros) and lesser kudu (T. imberbis) are even less closely related. In both cases, what we have are two species coincidentally developing a similar physical appearance, perhaps to adapt to similar conditions, and helped by the fact that, if they're not each others' closest relatives, they are at least close enough that they probably looked fairly similar to start with.

It's not unreasonable to suppose that the same thing happened to the bushbuck and imbabala, two animals so similar that they were long considered to be mere subspecies of the same creature. Case solved.

Or is it?

Since I last wrote about bushbuck, a new genetic study has been published. This showed that the bushbuck and imbabala are each other's closest relatives after all. While this doesn't prove that they were the same species all along - indeed, the study concluded that they probably weren't - it does at least make that possibility more likely. Perhaps more interesting, though, is the question of why they came to such a different conclusion as the earlier study, and which one happens to be right?

The answer to the first question is, as one might expect, that they used different methods. The 2008 study used mitochondrial DNA, which has the advantage that it is only ever inherited from mother to child. Thus, we can say that, if we trace the line back from mother to grandmother, and so on for however many thousands of generations it takes, we end up at two completely different places on the wider species family tree.

The 2018 study, however, used nuclear DNA, which has the advantage that there's a lot more of it, and it tends to give a more complete picture, including male ancestry as well as female. Using this, we find that we can get back to one single point on the family tree, a single ancestor from which both bushbuck and imbabala (and nothing else) descend.

Two different methods, two different answers. To resolve this conundrum, a third study has looked at both kinds of DNA, comparing the two species with each other, and with all the other spiral-horned antelopes, to try and produce the most accurate family tree yet. They were also able to use data from fossils of known age to determine not only what happened, but when.

The results confirmed the findings from the two earlier studies - both of them. In other words, the two sorts of DNA seem to have entirely different histories within the various Tragelaphus species. Combining this with various other sorts of statistical data, here's what they think happened:

Mitochondrial line in green
Numbers represent millions of years ago
Spiral-horned antelopes first evolved a little over 6 million years ago, towards the end of the Miocene. The oldest fossil examples are about this age and look rather like small nyala. The ancestors of the lesser kudu branched off not much later, while the remaining antelopes underwent two relatively rapid bursts of evolution in which new species were formed.

The first of these occurred during the mid-Pliocene at around 4 million years ago. Africa was becoming cooler at this time, as the world slipped into the long autumn that preceded the Ice Ages. Grasslands expanded at the expense of dense jungle, and we see the savannah-adapted greater kudu and elands (of which there are two species, which really are closely related) appearing at around this time, along with the nyala, which also prefers more open countryside.

The second burst of radiation occurs at around the dawn of the Pleistocene, just as the Ice Ages proper were getting started around 2 million years ago. It so happens that some pretty major geological changes were happening in East Africa at this time, mainly along the Rift Valleys. Combined with the wider changes to the worldwide climate, this sparked significant evolutionary changes in a number of animals, not least including our own ancestors.

This time saw the ancestors of the mountain nyala appearing, probably taking advantage of the cooler climate in the lowlands before such grassy terrain retreated to its present mountain home. Meanwhile, the ancestors of bongos sheltered in the few remaining patches of jungle, while the sitatungas rapidly adapted to live in waterways and swamps instead, so that the two now look quite different. The bushbucks also appear at this time, soon after splitting into two lineages - the present bushbuck and imbabala - perhaps separated from one another by the growing East African Rift Valley.

But something else, the researchers suggest, must have happened around the same time; the nyalas were also split into two species, living in different parts of the continent. One is the species we know today, while the other went extinct, living further north, but leaving no fossils we have yet found. No fossils but not, they argue, no descendants. Because at some point, a female of that lost species mated with a male bushbuck living in western Africa, and her descendants are the modern T. scriptus.

This is entirely possible. We know that even living species of spiral-horned antelope can crossbreed to produce fertile offspring, at least in the unnatural environment of zoos. While living bushbuck will, of course, have many ancestors that lived 2 million years ago, the single one in the direct female line - their Mitochondrial Eve - happens, if this is right, to be a member of an entirely different (and now extinct) species, related to the nyala.

The nuclear DNA reflects the bulk of the ancestors, the mitochondrial DNA follows just that one. This is termed introgression, where you have a single hybrid that then breeds with regular members of one of its ancestral species. Over the course of many generations, you have descendants that are mostly - but not entirely - members of one genetic species, with traces of the other buried deep in their genomes. It's the same reason that modern Europeans have some Neanderthal DNA - although, in our case, it doesn't happen to be the mitochondrial bit, which makes it much harder to spot.

Whether this process led, in this case, to the formation of an entirely new species, or whether bushbucks and imbabala are really just separate lineages within a single species, isn't something that can be so easily decided. Certainly, the known subspecies cluster together into two groups, one for the regular bushbuck, and one for the imbabala. So they are 'real', insofar as we can say that based on something that's inherently blurry.

They also appear to have different numbers of chromosomes, something that would normally mean that their offspring would be sterile, like mules, but which, due to some subtleties in their structure, doesn't. Is that enough to call them separate species, or is their interbreeding enough to say that they aren't?

That's a question we haven't yet unravelled.

[Photo by "Ossewa", from Wikimedia Commons. Cladogram adapted from Rakotoarivelo et al. 2019.]


  1. So, would it deserve a generic or subgeneric rearrange? Ammelaphus (imberbis), Nyala (angasi and extinct relative), Taurotragus (derbianus and oryx), Strepsiceros (strepsiceros), ? (buxtoni), Boocercus (euryceros and spekei) and Tragelaphus (scriptus and sylvaticus)

    1. While historically the elands have been given their own genus, I think this is difficult to maintain, and the introgression would make it hard to argue that the nyala and bushbuck aren't in the same genus. The lesser kudu is the one possible outlier (and would be Ammelaphus, as you say).