Selective feeding by a megaherbivore, the African elephant (Loxodonta africana)

1. What they did

Theory says the bigger a plant-eater is, and the more it relies on hindgut digestion, the less fussy it can afford to be. By that reasoning the African elephant should be the least selective herbivore of all. The authors point out that “selective” can mean choosing among plant species, or choosing among plant parts such as leaves, stems, bark and roots, and that these need to be kept apart.

They studied what elephants ate from trees and shrubs over one full year (August 1992 to July 1993) in the Kalahari sandveld savanna of Chobe National Park, northern Botswana, across three seasons: wet, cool dry and hot dry. They asked:

• Which woody species do elephants favour or reject in each season?
• Which species and plant parts make up most of what they eat?
• Which plant heights provide most of the food?
• What does this mean for how elephants change woodland?

Because there were too few roads to watch elephants feeding in the sandveld, they read what the elephants had eaten from the plants themselves. Early in the morning they found elephant tracks crossing roads and followed fresh feeding tracks away from the road. Along each track they sampled five circular patches of 10 m radius, 50 m apart, and checked every woody plant for fresh feeding, recording which parts had been eaten. Grass eating was not recorded.

2. What they found

Plant species

• In the wet season, only 8 of 27 common woody species (about 30%) were moderately or highly favoured. This rose to 10 species in the cool dry season and 15 (more than half) in the hot dry season.
• Diplorhynchus condylocarpon was the most favoured species in the cool dry season; Combretum fragrans and Combretum elaeagnoides stayed highly favoured throughout the dry season; Combretum zeyheri, Combretum collinum and Terminalia sericea became highly favoured only in the hot dry season.
• Some very common trees were rejected in every season: Baikiaea plurijuga, Ochna pulchra, Croton gratissimus and Erythrophleum africanum.
• Most of the browse eaten came from one or two common shrubs. In the wet season Baphia massaiensis and Bauhinia petersiana made up more than 60% of it; in the cool dry season Bauhinia petersiana and Diplorhynchus condylocarpon made up almost three-quarters; in the hot dry season Baphia massaiensis made up close to 40%.
• Almost all the highly favoured species were shrubs rarely taller than 5 m. Plants 1 to 2.5 m tall provided about 45% of the estimated diet in every season.

Plant parts

• Wet season: leaves 45% and shoots 25% of the woody part of the diet.
• Cool dry season: stems about 45%, leaves 24% and shoots 21%. Elephants stripped off and dropped the leaves of Diplorhynchus condylocarpon and several Combretum species, then ate the stems.
• Hot dry season: stems 55%, stem bark 25% and roots 14%; leaves and shoots only 1% each. Most roots came from plants under 2.5 m tall.
• No tree felling was seen at the sampled feeding sites. The damage was mainly leaf stripping and pruning, and the shrubs grew back strongly in the wet season.

3. What they concluded

• Elephants are narrowly selective among plant species, but extremely broad in the plant parts they eat. As food became scarce late in the dry season, they rejected almost no plant tissue.
• Rather than tolerating plant chemicals better than smaller browsers, elephants avoided the chemically rich leaves of many Combretum species that kudu, impala and giraffe readily eat. By turning to stems, bark and roots, they avoid chemicals found mainly in leaves and cope with the fibre through fast digestion.
• Because elephants ignore several common trees even as saplings, their feeding in this sandveld is more likely to change which species make up the woodland than to turn it into open grassland or shrubland. This may differ on richer soils, where Acacia trees are heavily used.
• Elephants’ reliance on bark, twigs and roots in the lean season sets them apart from all the ruminants they live with, more like rodents and hares and rabbits than like browsing antelope. This supports the idea that large herbivores divide up food mainly by plant part rather than by plant species.

4. How tracking was or was not used

Tracking was how the researchers found their samples. They looked for elephant tracks crossing roads in the early morning and judged them to be fresh feeding tracks if the plant parts dropped by the elephants had not yet dried out. They then followed these tracks away from the road, either the way the group had gone or back along where it had come from, for up to about 200 m. Along the feeding track they read the feeding signs on each woody plant: stripped leaves, broken shoots, eaten stems, bark eaten from stems and roots, eaten roots and eaten fruits.

5. Tracker or observer certification

Not reported. The paper does not say who found and followed the tracks or who judged the feeding signs, and gives no test of their skill. All the results depend on correctly telling fresh elephant feeding from older damage.

6. Other technologies, skills or innovations

• Plant-based observation: reading feeding signs on plants along a feeding track instead of watching the animals feed, useful where animals cannot be followed by vehicle.
• A random starting point for the first patch (a random number from a scientific calculator multiplied by 200 m), then patches spaced every 50 m.
• Acceptance frequencies, which compare how often a food was eaten with how often it was met along the animal’s path, avoiding some problems of comparing use with availability measured separately.

7. Implications for conservation (from the paper)

• In Kalahari sandveld, elephant feeding is more likely to change the mix of woody species than to wipe out the woodland, because several common trees are left alone even as saplings and the favoured shrubs regrow from large underground reserves.
• By eating small plants of favoured shrubs under 0.5 m, elephants may hold these shrubs back from growing taller.
• Elsewhere, the authors have seen local tree felling where fire or frost had reduced food in the shrub layer.
• Because elephants and browsing ruminants rely on different foods, especially in the dry season, the authors suggest that fewer browsing antelope where elephants are abundant may reflect this difference in diet rather than elephants directly outcompeting them.