Africa’s spiny trees offer lessons in understanding the earth’s ecology

Spines don’t stop animals from browsing, but slow down their feeding rates. Gareth Hempson

The Serengeti; Chobe; Kruger; the Masai Mara: in many people’s minds, savannas are Africa. This hasn’t always been true, though.

About 30 million years ago, the continent was largely covered by forests. So how did Africa’s vast, tree-dotted grassy ecosystems come to be?

Until now, fire has been viewed as the main protagonist. It was believed that blazes rolled back tree cover in the continent’s wetter regions and provided fire-promoting grasses with access to the light they need to thrive. This story played out around seven million years ago and is told in the fossil charcoal records and the DNA of fire-adapted savanna trees.

But research my colleagues and I have conducted reveals that savanna trees hold more secrets – and that medium-sized browsing animals played a much earlier role than fire in developing savannas.

We believe our research opens interesting new avenues for understanding how ecosystems evolve, particularly the role that animals have played and continue to play.

For example, our evidence that medium-sized browsing animals can have outsized effects on shaping ecosystems matches data from field-based studies. In these studies the animals’ intense browsing has been shown to create a “browse-trap” that prevents trees from growing out of the juvenile phase. This strongly limits their ability to reproduce, prevents them shading out grasses, and also makes them more vulnerable to fire.

And perhaps there’s a greater role for browsing animals to play in maintaining the world’s grazing lands as rising carbon dioxide levels promote woody encroachment of savannas and grasslands across the world.

Spiny trees

Spines – the sharp thorns of many savanna trees – are defences that protect tree canopies from being entirely consumed by animals. Spines don’t stop animals from browsing, but slow down their feeding so that it soon makes more sense for the browser to move on to a tree that is less well defended.

This means that the presence of lots of spiny trees is a good marker of high animal browsing pressure. It also turns out that spiny trees are particularly common in dry, nutrient-rich savannas where animals are abundant but fire is less common. We wondered whether the evolutionary histories of spiny trees like these could tell us something about the origins of these drier savannas.

So we set about analysing the evolutionary relationships of spiny trees in Africa. We did this by using DNA samples from 1,852 African tree species to reconstruct both when different species evolved and whether they had spines.

The analysis revealed a striking pattern. There were very few spiny tree species before about 16 million years ago. But what followed was an astonishingly rapid increase in the number of spiny species.

In fact, thorns evolved independently on 55 different occasions, with each event giving rise to a suite of spiny descendent species. More than 200 of these are still alive today. What triggered this sudden explosion of spine-defended trees in Africa?

Browsing antelope

Spines are a clear defence against being eaten, but not all browsing animal species are equally deterred by them. Elephants tend to shrug them off, while the smallest duiker species with their narrow muzzles are more adept at avoiding them. Spiny defences work best against medium to large browsing species like gazelles, impala and kudu.

This background knowledge set the scene for us. We knew which animals spines defend against, and that there was a sudden strong selection pressure favouring spiny tree species. All we needed was for the last piece of the puzzle to fall into place, which it did.

Africa was an island in the early parts of this story. It lacked the browsing animal species that spines defend against. But that all changed some 16 million years ago as the African and Eurasian continents drifted into one another. This formed a land bridge that allowed a novel suite of animals to venture into Africa.

A key group was the bovids. What followed was a remarkably convergent pattern of diversification in spiny tree species and also in the bovid animal groups that these spines defend against – and which appear to have changed the rules for persisting in woody communities.

Both of these plant and animal groups are now most diverse in the drier, nutrient-rich African savannas. This overwhelmingly suggests that browsing animals, not fire, engineered the first African savannas.


This research also poses a question to ecologists from across the globe: what is the legacy of animals in shaping ecosystems on other continents? The rest of the world suffered massive losses of large mammal species around 10,000 to 20,000 years ago. This was associated with the arrival of humans – a totally novel predator – and climatic fluctuation.

Are spiny vegetation types elsewhere a legacy of animals that are now extinct? How might those animals have engineered those environments? And can we use browsing animals to prevent trees from invading the world’s grazing lands, both in Africa and elsewhere?

Africa clearly has a crucial role to play in understanding the rest of the world’s ecology. It’s the only continent that still has relatively intact large mammal communities, and is the only place where we can really study how they’ve shaped the world we live in.

Author’s note: The research described in this article was led by Dr Tristan Charles-Dominique (University of Cape Town). The study emerged from a working group assembled by Dr Michelle van der Bank (University of Johannesburg), with key contributions from Professor William Bond (SAEON and the University of Cape Town) and Professor Jonathan Davies (McGill University).