Savanna plant survival: hanging out in the right crowd

Tyler Coverdale first visited the Mpala Research Centre in Laikipia, Kenya in 2013, and immediately became painfully aware of the abundant spiny and thorny plants that cover the savanna.  Spines help defend the plants from voracious elephants, giraffes and numerous other herbivores that depend on vegetation for their sustenance.


Camels browsing on  Barleria trispinosa at Mpala Research Centre, Kenya. Credit Tyler Coverdale.

Acacia trees such as Acacia etbaica (left foreground below) dominate the landscape, and may be associated with smaller shrubs, such as Barleria trispinosa. In the photo below, there is one B. trispinosa plant immediately below (on the right side) the acacia tree, and a second B. trispinosa plant to its right, more out in the open.  Coverdale realized that being situated immediately below a spiny acacia tree might be advantageous to B. trispinosa, which could be protected from the ravages of elephants and giraffes by the acacia thorns .

MRC landscape

Credit: Tyler Coverdale.

As you might guess by its name, B. trispinosa is itself a very spiny plant, which should help protect it from browsers.  Nonetheless, it still gets eaten, so Coverdale and his colleagues explored whether being under acacias would reduce how much it, and two other related species, got browsed.

Barleria trispinosa

Barleria trispinosa out in the open. Credit: Tyler Coverdale.

The first study was observational – a survey of the damage three species of Barleria suffered when they were under (associated with) acacia trees vs. unassociated with acacia trees. For each Barleria species, the researchers haphazardly chose 10 stems from eight associated and eight unassociated plants, and measured the proportion of these stems that showed physical evidence of being browsed.  As the figure below shows, browsing was sharply lower for each species when it was associated with an acacia plant.


Percentage of stems damaged by browsers for three Barleria species in relation to whether they were associated or unassociated with an acacia tree.* indicates significant differences between means in all figures.

The understory plant community associated with acacias is much denser than the plant community out in the open, so the researchers wondered whether it was the acacia itself, or the other plants associated with it, that were providing protection. They set up an experiment using focal B. trispinosa plants with four treatments (A) unmanipulated control, (B) overstory removal, (C) overstory + understory removal, (D) a procedural control with overstory + understory removal, with the focal plant enclosed in a metal cage to protect it from predators (see Figure below).


Coverdale and his colleagues ran the experiment for one month.  They discovered that removing overhanging acacia branches sharply increased herbivory, but the additional removal of understory neighbors had little additional effect.  Both the unmanipulated controls and procedural controls were unaffected.


Change in % of stems browsed for (A) unmanipulated control (left bar), (B) overstory removal (second from left bar), (C) overstory + understory removal (second from right bar), (D) a procedural control (right bar).  Different letters above bars indicate significant differences between the mean values.

The researchers then investigated how useful these spines are to unassociated B. trispinosa plants. They set up another experiment with four types of spine treatments: (A) unmanipulated controls, (B) 50% spine removal, (C) 100% spine removal, (D) procedural control with 100% spine removal + enclosure within a predator-proof cage. These cages were vandalized shortly after the experiment was set up, so the researchers chose eight plants from a nearby plot (that had all predators excluded for a different experiment) as their procedural control. They discovered that spines are very useful to protect against predators in unassociated B. trispinosa.


Change in % of stems browsed for (A) unmanipulated control (left bar), (B) 50% spine removal (second from left bar), (C) all spines removed (second from right bar), (D) procedural control (right bar).

If you were a plant living under the protection of an acacia tree, it would make sense for you to reduce your investment in thorns, so you could allocate more resources to growth and reproduction.  Does Barleria do this?


Several lines of evidence indicate that all three Barleria species reduce their investment in spines when associated with an acacia. First, a survey of spine density shows a reduced number of spines for all three species when they were associated with acacia trees (top graph).  Second, the spines that are present are significantly shorter in Barleria species associated with acacia trees (middle graph).  In a final survey, Coverdale and his colleagues cut all of the spines off of associated and unassociated Barleria.  For each plant, the researchers calculated the dry weight of spines and of all the other plant tissue.  For each Barleria species, the defensive investment – the ratio of spines to total mass, was substantially reduced in acacia-associated plants in comparison to unassociated plants (bottom graph).

Lastly, can plants react adaptively to browsing?  In other words, will understory plants produce more thorns if they are browsed?  To explore this question, the researchers used scissors to simulate moderate (25%) or heavy (50%) browsing.  They discovered a significant increase in spines produced by unassociated plants one month after clipping. Ecologists call this an induced defense. This induced defense is strongly suppressed in plants that have lived under the protection of acacia trees – in fact there was no significant response to experimental browsing in acacia-associated B. trispinosa plants. The researchers don’t know how long this suppression of induced responses persists. Would browsing induce increased spine growth in B. trispinosa six months, a year or two years after its protective acacia tree died?

Coverdale and his colleagues conclude that the overall benefit of association is positive to the plant populations.  Their studies show better survival and higher reproductive rates of acacia-associated understory plants. There is probably a cost associated with too many offspring competing for resources within a small area, as seedlings tend to grow within 1 meter of their parents.  However the reduction in defense costs probably overrides this cost of competition, leading to increased population size.  The researchers suggest a long-term study of population growth rates for acacia-associated and unassociated plants for several different species to see how general these effects are, and to explore whether other factors, such as soil moisture and nutrient levels influence the allocation and induction of defensive structures such as spines and thorns.

note: the paper that describes this research is from the journal Ecology. The reference is Coverdale, T. C., Goheen, J. R., Palmer, T. M. and Pringle, R. M. (2018), Good neighbors make good defenses: associational refuges reduce defense investment in African savanna plants. Ecology, 99: 1724-1736. doi:10.1002/ecy.2397. Thanks to the Ecological Society of America for allowing me to use figures from the paper. Copyright © 2018 by the Ecological Society of America. All rights reserved.

Fires foster biological diversity on the African savanna

As an ecology student back in days of yore, I was introduced to the classic mutualism between ants and swollen-thorn acacia trees. In this mutually beneficial relationship, ants protect acacia trees by biting and projecting very smelly substances at hungry herbivores, and by pruning encroaching branches of plant competitors. In return for these services, acacia trees provide the ants with homes in the form of swollen thorns, and in some cases also provide food for their defenders.


Swollen thorns of Acacia drepanlobium occupied by C. nigriceps. Credit: Ryan L. Sensenig.

I always assumed there were limits to what these ants could do. I knew that elephants were a constant problem for trees trying to get established on the African savanna. I figured, wrongly, that ants could not do much to counter a determined thick-skinned elephant. But as Ryan Sensenig describes, ants will swarm any intruding elephant, rushing into the elephant’s very sensitive trunk and mouth, biting it and, in some cases, exuding a chemical compound that is very offensive to an elephant’s keen sense of smell. So don’t mess with these ants if you can help it!


The Laikipia Plateau has one of the few growing elephant populations in East Africa. Credit: Ryan L. Sensenig.

Fires play an important role in savanna ecosystems, killing many trees before they can get established, and creating a mosaic of burned and unburned areas which vary in grass quality and quantity, and in the abundance of acacia trees (and other species as well). Recently burned grasslands tend to be lower in grass abundance and higher in grass nutrient levels. In a previous study of controlled burns, Sensenig and his colleagues showed that larger animals, such as elephants, tended to graze in unburned areas, which had more grass – albeit of lower quality. Returning seven years after the burn, he was surprised to find that elephants, which eat both trees and grass, had shifted to the burned sites in preference to unburned sites. He thus wondered whether fire was having an impact on the ant-acacia mutualisms that defend acacias from elephants and other large herbivores.


Sunset strikes an Acacia xanthophloea on Mpala Research Centre in Laikipia, Kenya. Credit: Ryan L. Sensenig.

Ants do not share trees. In Mpala Research Centre in the Laikipia Plateau of Kenya, there are four mutually-exclusive species of ants that live in Acacia drepanolobium trees: Crematogaster sjostedti, C. mimosae, C. nigriceps, and Tetraponera penzigi.

Sensenig and his colleagues wanted to know whether the controlled burns had a long-lasting effect on ant species distribution on acacia trees. The researchers surveyed 12 plots that had been burned seven years previously and an equal number of unburned plots to see how burns affected which ant species were present.


Goshen College research students estimate ant densities on Acacia drepanolobium trees in the Kenya Longterm Exclosure Experiment. Credit: Ryan L. Sensenig.

They found that C. nigriceps was more common in acacias from burned areas while the other three species were more common in trees from unburned areas.


Why were there more C. nigriceps ants in previously burned areas? One explanation is that perhaps C. nigriceps is better at avoiding getting burned by fire, or else is better at recolonizing after a fire. To look for species difference in response to fire, the researchers simulated fires by burning elephant dung and dried grass in 3-gallon metal buckets, creating a small sustained smoke source. They stationed observers every 50 meters along a 500 meter transect for the first experiment, and a 1.8 km transect for the second experiment. They then measured ant-evacuation rate by counting the number of ants moving down the trunk. There were some very pronounced differences, with C. nigriceps having the highest evacuation rate, C. mimosae also showing a strong smoke response, and the other two species showing little evidence of any response.


Evacuation rate for each species in response to smoke.

C. mimosae generally prevails when it battles a colony of C. nigriceps. These results indicate that the subordinate C. nigriceps is able to maintain its presence in the community, in part, by taking advantage of its superior performance when it encounters a fire. The researchers also found some evidence that C. nigriceps is better at recolonizing after a fire than is C. mimosae. So despite being the subordinate species, C. nigriceps is abundant in this ecosystem.

Returning to those elephants, the researchers describe one final experiment in which some plots had a series of fences that excluded herbivores, while other plots were open to herbivores, including elephants.


In this experiment, as well, there were burned and unburned plots. In general, there were more ants present when herbivores were present, as the trees invested more in swollen thorns and in ant food (in the form of nectar) to attract protective ants. In addition, ants were more abundant in unburned plots than in plots that had been previously burned, with the exception of C. nigriceps when herbivores were excluded.

Ecologists have long known that fire maintains savanna ecosystems by preventing the grasslands from being overgrown by trees. This study shows that fires shift ant community structure in favor of the subordinate ant species (C. nigriceps), resulting in a higher diversity of ant species overall. The researchers suggest that if fires become more common in savannas, elephants may become more attracted to acacias that harbor a reduced (or nonexistent) cast of defenders, which could lead to a further reduction in the abundance of acacia trees and their mutualistic ants.

note: the paper that describes this research is from the journal Ecology. The reference is Sensenig, R. L., Kimuyu, D. K., Ruiz Guajardo, J. C., Veblen, K. E., Riginos, C., & Young, T. P. (2017). Fire disturbance disrupts an acacia ant–plant mutualism in favor of a subordinate ant species. Ecology, 98(5), 1455-1464.Thanks to the Ecological Society of America for allowing me to use figures from the paper. Copyright © 2017 by the Ecological Society of America. All rights reserved.