The Geographical Pattern of Ebola Outbreaks

Since 1976 Ebola virus has hopscotched its way around Central Africa. What’s going on?

The geographical pattern of Ebola outbreaks among humans is controversial. Everyone knows what that pattern looks like but experts dispute what it means. The dispute involves Ebola virus in particular, the one among those five ebolaviruses that has emerged most frequently, in multiple locations across Africa, and therefore cries most loudly for explanation. From its first known appearance to the present, from Yambuku (1976) to Tandala (1977) to the upper Ivindo River gold camps (1994) to Kikwit (1995) to Mayibout 2 (1996) to Booué (later 1996) to the northern border region between Gabon and the Republic of the Congo (2001–2002) to the Mbomo area (2002–2003) to its recurrence at Mbomo (2005) and then to its two more recent appearances near the Kasai River in what’s now the Democratic Republic of the Congo (2007–2009), Ebola virus has seemingly hopscotched its way around Central Africa. What’s going on? Is that pattern random or does it have causes? If it has causes, what are they?

Two schools of thought have arisen. I think of them as the wave school and the particle school—my little parody of the classic wave-or-particle conundrum about the nature of light. Back in the seventeenth century, as your keen memory for high-school physics will tell you, Christiaan Huygens proposed that light consists of waves, whereas Isaac Newton argued that light is particulate. They each had some experimental grounds for believing as they did. It took quantum mechanics, more than two centuries later, to explain that wave-versus-particle is not a resolvable dichotomy but an inef- fable duality, or at least an artifact of the limitations of different modes of observing.

The particle view of Ebola sees it as a relatively old and ubiquitous virus in Central African forests, and each human outbreak as an independent event, primarily explicable by an immediate cause. For instance: Somebody scavenges an infected chimpanzee carcass; the carcass is infected because the chimp itself scavenged a piece of fruit previously gnawed by a reservoir host. The subsequent outbreak among humans results from a local, accidental event, each outbreak therefore representing a particle, discrete from others. Eric M. Leroy, a Paris-trained veterinarian and virologist, is the leading proponent of this view. “I think the virus is present all the time, within reservoir species,” he told me. “And sometimes there is transmission from reservoir species to other species.”

Spillover: Animal Infections and the Next Human Pandemic by David Quammen (W. W. Norton & Company, 2012)

The wave view suggests that Ebola has not been present throughout Central Africa for a long time—that, on the contrary, it’s a rather new virus, descended from some viral ancestor, perhaps in the Yambuku area, and come lately to other sites where it has emerged. The local outbreaks are not independent events, but connected as part of a wave phenomenon. The virus has been expanding its range within recent decades, infecting new populations of reservoir in new places. Each outbreak, by this view, represents a local event primarily explicable by a larger cause—the arrival of the wave. The main proponent of the wave idea is Peter D. Walsh, an American ecologist who has worked often in Central Africa and specializes in mathematical theory about ecological facts.

“I think it’s spreading from host to host in a reservoir host,” Walsh said, when I asked him to explain where the virus was traveling and how. This was another conversation in Libreville, a teeming Gabonese city with pockets of quietude, through which all Ebola researchers eventually pass. “Probably a reservoir host that’s got large population sizes and doesn’t move very much. At least, it doesn’t transmit the virus very far.” Walsh didn’t claim to know the identity of that reservoir, but it had to be some animal that’s abundant and relatively sedentary. A rodent? A small bird? A nonmigrating bat?

The evidence on each side of this dichotomy is varied and intriguing, though inconclusive. One form of that evidence is the genetic differences among variants of Ebola virus as they have been found, or left traces of themselves, in human victims, gorillas, and other animals sampled at different times and places. Ebola virus in general seems to mutate at a rate comparable to other RNA viruses (which means relatively quickly), and the amount of variation detectable between one strain of Ebola virus and another can be a very important clue about their origins in space and time. Peter Walsh, working with two coauthors on a paper published in 2005 (“Wave-Like Spread of Ebola Zaire”), combined such genetic data with geographical analysis to suggest that all known variants of Ebola virus descended from an ancestor closely resembling the Yambuku virus of 1976.

Maps of Ebola Zaire Outbreaks [Walsh, Real, Biek]

Walsh’s collaborators were Leslie Real, a highly respected disease ecologist and theoretician at Emory University, and a bright younger colleague named Roman Biek. Together they presented maps, graphs, and family trees illustrating strong correlations among three kinds of distance: distance in miles from Yambuku, distance in time from that 1976 event, and distance in genetic differences from the Yambuku-like common ancestor. “Taken together, our results clearly point to the conclusion that [Ebola virus] has gradually spread across central Africa from an origin near Yambuku in the mid-1970s,” they wrote. Their headline, stating the thesis plainly, was WAVE-LIKE SPREAD OF EBOLA ZAIRE. It may or may not be a new pathogen—at least, new in these places. (Other evidence, published more recently, suggests that filoviruses may be millions of years old.) But maybe something happened, and happened rather recently, to reshape the virus and unleash it upon humans and apes. “Under this scenario, the distinct phylogenetic tree structure, the strong correlation between outbreak date and distance from Yambuku, and the correlation between genetic and geographic distances can be interpreted as the outcome of a consistently moving wave of [Ebola virus] infection.” One consequence of the moving wave, they argued, is massive mortality among the apes. Some regional populations have been virtually exterminated—such as the gorillas of the Minkébé forest, of the Lossi sanctuary, of the area around Moba Bai—because Ebola hit them like a tsunami.

So much for the wave hypothesis. The particle hypothesis embraces much of the same data, construed differently, to arrive at a vision of independent spillovers, not a traveling wave. Eric Leroy’s group also collected more data, including samples of muscle and bone from gorillas, chimps, and duikers found dead near human outbreak sites. In some of the carcasses (especially the gorillas), they detected evidence of Ebola virus infection, with small but significant genetic differences in the virus among individual animals. Likewise they looked at a number of human samples, from the outbreaks in Gabon and the Congo during 2001–2003, and identified eight different viral variants. (These were lesser degrees of difference than the gaps among the five ebolaviruses.) Such distinct viruses, they proposed, should be understood in the context that their genetic character is relatively stable. The differences among variants suggest long isolation in separate locales, not a rolling wave of newly arrived, rather uniform virus. “Thus, Ebola outbreaks probably do not occur as a single outbreak spreading throughout the Congo basin as others have proposed,” Leroy’s team wrote, alluding pointedly to Walsh’s hypothesis, “but are due to multiple episodic infection of great apes from the reservoir.”

This apparent contradiction between Leroy’s particle hypothesis and Walsh’s wave hypothesis reflects an argument at cross-purposes, I think. The confusion may have arisen from back-channel communications and a certain sense of competition as much as from ambiguity in their published papers. What Walsh suggested—to recapitulate in simplest form—is a wave of Ebola virus sweeping across Central Africa by newly infecting some reservoir host or hosts. From its recent establishment in the host, according to Walsh, the virus spilled over, here and there, into ape and human populations. The result of that process is manifest as a sequence of human outbreaks coinciding with clusters of dead chimps and gorillas—almost as though the virus were sweeping through ape populations across Central Africa. Walsh insisted during our Libreville chat, though, that he had never proposed a continental wave of dying gorillas, one group infecting another. His wave of Ebola, he explained, has been traveling mainly through the reservoir populations, not through the apes. Ape deaths have been numerous and widespread, yes, and to some degree amplified by ape-to-ape contagion, but the larger pattern reflects progressive viral establishment in some other group of animals, still unidentified, with which apes frequently come into contact. Leroy, on the other hand, has presented his particle hypothesis of “multiple independent introductions” as a diametric alternative not to Walsh’s idea as here stated but to the notion of a continuous wave among the apes.

In other words, one has cried: Apples! The other has replied: Not oranges, no! Either might be right, or not, but in any case their arguments don’t quite meet nose to nose.

So . . . is light a wave or a particle? The coy, modern, quantum-mechanical answer is yes. And is Peter Walsh correct about Ebola virus or is Eric Leroy? The best answer again may be yes. Walsh and Leroy eventually coauthored a paper, along with Roman Biek and Les Real as deft reconcilers, offering a logical amalgam of their respective views on the family tree of Ebola virus variants (all descended from Yambuku) and of the hammer-headed bat and those two other kinds of bats as (relatively new) reservoir hosts. But even that paper left certain questions unanswered, including this one: If the bats have just recently become infected with Ebola virus, why don’t they suffer symptoms?

The four coauthors did agree on a couple other basic points. First, fruit bats might be reservoirs of Ebola virus but not necessarily the only reservoirs. Maybe another animal is involved— a more ancient reservoir, long since adapted to the virus. (If so, where is that creature hiding?) Second, they agreed that too many people have died of Ebola virus disease, but not nearly so many people as gorillas.

You have been reading an excerpt from Spillover: Animal Infections and the Next Human Pandemic by David Quammen, available wherever books are sold:

Amazon | Barnes & Noble | iBookstore | Indiebound | Powell’s

DAVID QUAMMEN is the author of The Song of the Dodo, among other books. He has been honored with the John Burroughs Medal for nature writing, an the Academy Award in Literature from the American Academy of Arts and Letters, an award in the art of the essay from PEN, and (three times) the National Magazine Award. Quammen is also a contributing writer for National Geographic. He lives in Bozeman, Montana. Learn more at and find him on Twitter at @DavidQuammen.