A person has time to think when sitting in a parked car, waiting for someone. I often found myself in this situation while studying the feeding patterns of giraffes, kudus, impalas, and steenbok in South Africa’s Kruger National Park. On a typical day, lunchtime music was playing softly on my radio, and cicadas shrilled in the grass. The landscape was dry and dusty, except for the rivers in the distance, which were visible as bands of green vegetation that stood in bold contrast to the brown hills. I might have been watching one of my study animals, perhaps a giraffe cow ruminating in the shade of a knobthorn acacia. Beneath a veil of creamy white flowers, the gray limbs of the tree shimmered in the midday heat. The giraffe, her ears laid back and her eyes half shut, chewed steadily on acacia flowers.
This scene is typical of the late dry season in the lowveld, South Africa’s expanse of semiarid savanna. The lowveld receives only twenty-four inches of rainfall each year, mostly in the austral midsummer from November to February. During this short wet season, plants sprout new leaves and many burst into flower. Antelopes and many other medium-sized herbivores calve at this time to take advantage of the plentiful food. With the onset of the dry season, however, conditions begin to deteriorate as soil moisture evaporates and seeps into the folds of the landscape. By August, most trees are dormant, standing bare over the brown earth and brittle grass, and most large herbivores have moved downslope to riverbanks where they can still find some green leaves to browse on. The animals must survive several more hot, dry months before purple-gray rain clouds arrive. Ribs and pelvic bones will protrude farther before the green flush of the next rainy season.
During my three years in Kruger National Park, I spent my days tracking the movements of my study animals and recording their choices of habitats and food. Many interesting results emerged as my work proceeded, but perhaps the most remarkable was the relationship between giraffes and the knobthorn acacias (Acacia nigrescens), common trees of the dry plains and ridges. Toward the end of the dry season, giraffes are lured away from the riverine vegetation by the knobthorn flowers. Borne in clusters on the ends of branches, knobthorn blossoms are easily cropped by giraffes as they roam from tree to tree, and provide them with up to a quarter of their food during August and September, when the trees are in full bloom. Other browsers lack the giraffes’ height advantage and are unable to reach flowers on most knobthorn trees, although I often saw impalas scavenging dead flowers that had fallen to the ground.
At various times of the year, I observed giraffes feeding on the flowers of other tree species, but these provide little food compared with the knobthorns. I found it odd that only the knobthorns produce such an abundance of flowers at a time when the giraffes need them most as food. In the dry season, the survival of plants depends on adaptations that protect them from browsers. So why hadn’t the knobthorns evolved defenses to protect their flowers? Wouldn’t the loss of so many flowers reduce the trees’ ability to reproduce?
As my research progressed, I began to wonder if the relationship between giraffes and flowering knobthorns might not be as one-sided as it first appeared. Perhaps the giraffes were paying for a palatable meal by carrying pollen between patches of flowering knobthorns. This possibility became increasingly plausible as I learned more about the reproductive biology of African acacias.
[media:node/690 large vertical caption]
Although the pollinators for only a few species of African acacia have been identified, the general assumption has been that insects perform the task. (The pollen grains of acacia trees are too large and heavy to be effectively dispersed on the wind. Birds and bats are unlikely pollinators, because the knobthorns do not provide the nectar they seek.) Therefore, on the plains of a semiarid savanna, it makes sense for a tree to flower in the wet season, especially after each rainstorm, when flying insects that are likely to transfer pollen are most active. The umbrella thorn (Acacia tortilis), one of the most common trees in my study area, follows this pattern, flowering sporadically in the wet season. Each flowering is triggered by recent rain so that the same tree may flower several times a season.
During the hot dry season, however, bees are rarely found far from water; I often saw swarms of them along the edges of stagnant pools or on damp sand in drying riverbeds. It makes sense for riverine trees to flower in the dry season when competition for insect pollinators is low and bee activity is restricted mainly to the same habitat as the flowering plants. Indeed, Kipling’s fabled fever trees (A. xanthophloea) occur only along rivers and bloom in the late dry season, when their golden "pompon" flowers sweeten the air and the riverine canopy hums with bees.
The distribution of flowering knobthorns, however, does not appear to make much sense. Although they also flower late in the dry season, they are rarely found along rivers and are often miles from surface water—and bees. On the hot, arid plains, their flowering canopies are striking to the eye, but they are neither fragrant nor humming.
Based on the general appearance of their flowers, taxonomists divide the African acacias into two different groups. One, which includes the umbrella thorn and the fever tree, generally has bright golden or orange-yellow pompon flower clusters and is probably pollinated by insects. These acacias flower at times when flying insects are active in their habitats, their fragrance and colors attract insects, and their spiny thorns protect their flowers from browsing herbivores. The other group, which includes the knobthorn, has flower clusters that resemble slender bottle brushes and do not seem to be as well adapted for insect pollination. In fact, the knobthorn’s pale flowers fit the tendency for flowers pollinated by mammals to have dull colors.
Bats are the most common mammalian pollinators, but within the last decade, a number of small, flightless mammals have also been found to transfer pollen. For instance, Delbert Wiens, a biologist at the University of Utah, found that the flowers of the South African protea are pollinated by rodents (see "Secrets of a Cryptic Flower," Natural History, May 1985). Mammals are covered with fur or hair, which is good for collecting and carrying pollen grains, and because they are so much larger than insects, they can carry more. Their biggest disadvantage is that most mammals cannot fly, making them poor cross-pollinators. Efficient cross-pollinators must have easy access to flowers in the tree canopy and the ability to travel directly between widely distributed plants of the same species. Giraffes meet all these qualifications. While browsing fifteen feet above the ground, their heads and necks become covered with pollen as they shake branches and brush through clusters of blossoms. (While inspecting trees on which the giraffes had fed, I inadvertently collected pollen on my forearms.) Giraffes are capable of traveling more than ten miles a day between stands of flowering trees.
All plants pollinated by animals have to pay a price for the service, usually in the form of pollen and nectar, which are rich in energy and protein. Because giraffes operate on a larger scale, their reward is whole flowers—the one disadvantage to the tree of giraffe pollination. Their potential as pollen vectors remains academic unless the benefits of pollen transfer override the drawbacks of flower loss.
Browsing by giraffes is less of a threat to the reproductive fitness of flowering knobthorns than might be imagined, however. Despite swarms of flying insects around the flowers of some species, African acacias are noted for a particularly low ratio of seed pods to flowers. Many of the flowers may never get fertilized, but another reason for the small number of pods may be that most, and possibly all, African acacia flower clusters are mixtures of bisexual, male only, and sterile flowers. Some botanists have suggested that large numbers of male and sterile flowers could function as petals and are simply there to attract pollinators. The loss of sterile flowers to giraffe browsing is of no consequence to the reproductive fitness of the acacia trees. Nor is the consumption of male flowers if their pollen is transferred in the process. Furthermore, since all the knobthorns produce flowers simultaneously, the giraffe population (which is limited by the scarcity of food during the rest of the dry season) cannot consume a significant fraction of the female flowers in the short time that the trees bloom.
Even though an animal may be capable of transferring pollen, it is not necessarily a reliable pollinator. The animal must visit the plants regularly, as part of its usual feeding pattern. In central Kruger, this is precisely the case. Knobthorns provide giraffes with much more than flowers. During the wet season, knobthorn foliage is the staple diet of giraffes, making up more than 40 percent of their annual diet. The intimate link between the two species is reflected in the distribution of giraffes and knobthorns in the park, which can be neatly superimposed upon a map.
Perhaps the strongest support for my suggestion that giraffes pollinate knobthorns is the trees’ lack of defenses. If giraffes browsing on flowers were harmful to the reproductive success of knobthorns, there would be selective pressure for the evolution of an effective defense. Witness the well-defended pompon flowers of the umbrella thorn, a species that is pollinated by insects. Borne on short stalks, the flower clusters lie close to the branches, between long and formidable thorns. In contrast, the knobthorn’s bottle-brush flowers extend beyond the protection of the trees’ prickles. Nor are the knobthorn flowers defended by noxious chemicals. In contrast, I never saw a giraffe even nibble at the spectacular bright yellow flowers of a Cassia abbreviata, a tree related to acacias, even though they are easily accessible, suggesting that they may contain a particularly unpleasant toxin.
While I still lack definitive proof that giraffes are pollinating knobthorns, there is solid evidence for the coevolution of some acacias and large African herbivores. The seedpods of the umbrella thorn are aromatic and nutritious, so when they are shed in the dry season they are avidly sought out by impalas, kudus, steenbok, and other herbivores. Their hard, thick coats protect the seeds from damage as they pass through the herbivore’s gut, and the seeds are ready for germination when they are eventually deposited in dung some distance from the parent plant. If coevolution between large herbivores and African acacias could result in an efficient system of seed dispersal, I see no reason why it could not lead to efficient pollen dispersal as well.
Traditionally, pollination by animals has been regarded as the exclusive domain of small species that fly. Only within the last decade have some small flightless mammals proved to be pollen vectors, and these discoveries were initially considered bizarre. Now I am really bucking tradition by suggesting that some acacias may be pollinated by the tallest living animal. I must point out, however, that this story represents only the first step of scientific inquiry: the formulation of a testable hypothesis. My next step will be to test whether making knobthorns inaccessible to giraffes will significantly reduce production of knobthorn seeds. This will not be easy, however, because any barriers I erect to exclude giraffes must be strong enough to avoid destruction by elephants, which can wreck just about anything--especially fences. But the effort will be worthwhile if it confirms that the knobthorn has evolved a pollination system that relies on the world’s tallest animal.