The Monkey's Voyage

by Alan de Queiroz

  8.4 Ptychadena newtoni, one of the endemic amphibians of São Tomé and Príncipe, and the subject of Measey et al.’s key 2007 biogeographic study. Photo by Andrew Stanbridge.

  After getting more DNA sequences from the tissue samples, the group used the usual computer programs to crank out an evolutionary tree of the genus Ptychadena. Sure enough, the specimens of P. newtoni, from São Tomé, were all together on their own branch, well separated genetically from any other species. So, just as with the frogs on Mayotte, P. newtoni is apparently a unique island species and therefore likely arrived on São Tomé by a natural overwater crossing long before there were any boats to carry frogs there. This conclusion wasn’t too surprising, because Measey and the others already knew that P. newtoni was anatomically different from all other species of Ptychadena.

  The really odd thing about the Ptychadena results was where the P. newtoni branch arose in the evolutionary tree. You would think that the nearest relatives of P. newtoni would be from the Gulf of Guinea coast, because that location would require the shortest journey to São Tomé. As a parallel, the closest relatives of the Baja California garter snakes that Robin Lawson and I studied turned out to be almost directly across the Sea of Cortés from the peninsular populations. That is not the case, though, for P. newtoni. Instead, Measey and colleagues found that its closest relative is an undescribed species from Kenya, Tanzania, Uganda, and Egypt, that is, most of the way across the continent in East Africa. The next closest related frogs in the evolutionary tree are from West Africa, so it’s plausible that the ancestor of both P. newtoni and the undescribed species was found in West Africa, and that those two lineages dispersed independently to their current ranges. Still, that undescribed East African relative raises the possibility that the ancestors of P. newtoni instead made an epic journey from the eastern part of the Congo Basin many hundreds of miles downriver to the Atlantic, and from there were carried by ocean currents to São Tomé (see Figure 8.5).

  8.5 A possible route (dotted line) for rafting amphibians from the mouth of the Congo to São Tomé. The white arrow indicates the direction of the Congo Current. Redrawn and modified from Measey et al. (2007).

  Forced to choose between those two possibilities, I would normally choose the first one, the one that does not require an astounding voyage down the Congo. But that would be forgetting the cobra bobo. As Measey had mentioned to the oceanographer, Alain Morlière, the closest relatives of the cobra bobo are in Central and East Africa, not West Africa. This eastern connection could just be a coincidence, and a reflection of the fact that there are many amphibian species still to be discovered in Africa; maybe some as-yet-unknown West African frog and caecilian will turn out to be the actual closest relatives of P. newtoni and the cobra bobo. However, two of the other frogs of São Tomé and Príncipe also have their closest known relatives in East Africa. That means that, to the best of our knowledge, four of the five amphibian lineages on São Tomé and Príncipe (representing the five presumed colonizations from the mainland) are most closely related to species that live in the eastern half of the continent, all either in or close to the drainage of the Congo River. (The fifth lineage has not been studied, so it is unknown whether its affinities also lie in the eastern Congo Basin.) One or two species with this eastern connection could just be a fluke arising from our incomplete knowledge of African amphibians, but four must be more than a coincidence. It means that we should look for an explanation of how amphibians dispersed from East Africa to São Tomé and Príncipe.

  For skeptical biogeographers, the example of the Gulf of Guinea amphibians may bring to mind Gary Nelson’s contemptuous phrase that dispersalism is “a science of the improbable, the rare, the mysterious, and the miraculous.” Nelson’s words were wrapped up in his search for generality and his adherence to Karl Popper’s notion of falsification, but, in another sense, they were more transparent and commonsensical: he was pointing out the absurdity of many dispersal hypotheses, such as, say, amphibians traveling on a giant raft down the Congo and up the west coast of Africa to colonize São Tomé and Príncipe. I have to admit that these stories sometimes do sound ridiculous. However, as I have learned more and more about biogeography, a word for strange dispersal explanations—other than improbable, rare, mysterious, and miraculous—often comes to mind. The connection between the Gulf of Guinea amphibians and East African species makes me think of that word. The word is necessary.

  With the Ptychadena project in full swing, John Measey finally called up Bernard Bourles, one of the oceanographers who had intimate knowledge of the plume of water emanating from the Congo. What Bourles knew about the Congo plume was even better than Measey had hoped. It was the last big piece of the puzzle.

  A small river can create an area of brackish water around its mouth, but a really big river can do much more than that. Fresh water tends to float on top of the denser salt water of the ocean, and a river the size of the Congo can alter the salinity at the surface of the ocean hundreds of miles from the river’s mouth. In the rainy season, the combined output of the Congo, south of the Gulf of Guinea, and the Niger, which empties into the gulf itself, substantially lowers the salinity at the surface of the ocean from the Congo’s mouth through the entire gulf. That was the data that Bourles had, showing arcs of lowered salinity radiating out from the mouths of those two big rivers. What’s more, at various times, both historically and prehistorically, rainfall in tropical Africa has been much greater than it is today, which means more water would have poured out of the Congo, and the “freshening” of surface seawater would have been even stronger. In short, there have been times when the surface waters from the Congo’s mouth all the way to São Tomé and Príncipe probably have been heavily diluted with fresh water.

  Amphibians crossing ocean barriers need all the help they can get, and it may be that being carried on a natural raft floating on brackish water, rather than seawater, was the last piece that made the voyage from the Congo to São Tomé and Príncipe doable. Waves might be crashing over the raft, and, in any case, water would be seeping constantly into the matrix of earth and plant roots, even on a calm sea. In those circumstances, it would be far better for a frog or a cobra bobo if that water was much less salty than seawater. This is what Measey and his collaborators were arguing.

  Let us now imagine the whole story, with some help from images that Bob Drewes has collected. It’s some time within the past few million years, and it’s the rainy season in the upper eastern reaches of the Congo River Basin. The banks of the Congo are soaked with rain, weakened to the point that big chunks, complete with tall forest trees and undergrowth, are breaking off and sliding into the river. Most of the chunks break up without getting very far, but some of them, maybe the biggest ones, stay intact, or at least partially intact, and float far down the river. It’s not a stretch to think that most of these natural rafts have some amphibians on them, lodged in a burrow or among the roots of a tree—that would be the case for almost any sizable chunk of African rainforest.

  In places, the Congo forms rapids, and it passes over several waterfalls. Most of the rafts don’t make it past those gauntlets, but a few bump along through the whitewater stretches and crash over the falls, losing some large pieces and a few unhappy frogs in the process, but remaining mostly intact. After that, it’s a long drift down to the ocean.

  To make the point that very large natural river rafts are a reality, Bob Drewes points to aerial photos of the Amazon’s mouth, where giant rafts called camalotes have run aground in the delta. In one case, the camalotes have piled up over the years to form an island roughly the size of Belgium. The Amazon delta actually is not a great place for a large raft to begin an ocean journey, because the shallows are vast and most rafts get stuck, but Amazon camalotes occasionally do make it out to open water. A raft reaching the mouth of the Congo is much more likely than one in the Amazon to reach the open ocean, because the shallows there are not so extensive.


  Once a raft makes it offshore, the prevailing Congo Current will push it northward toward the Gulf of Guinea. At this point, for a frog or caecilian on a raft, the odds of making it to São Tomé or Príncipe are still not looking very good. The raft is traveling in the general direction of the Gulf of Guinea islands, but it still has about six hundred miles to go, miles in which it could easily break up or become so inundated that even the brackish water produced by the Congo plume will dehydrate any amphibian beyond recovery. Even most rafts that stay intact will not bring live amphibians to the islands. Bourles, the oceanographer, estimated that under perfect conditions, a raft might drift from the mouth of the Congo to São Tomé or Príncipe in two weeks, but it might take considerably longer. Some rafts will take too long to reach the islands, and the animals will starve or desiccate. Most of the rafts, like ships without pilots, will simply miss making landfall on the islands altogether. The ocean clearly acts as a filter that some terrestrial organisms have a hard time passing through. Almost all amphibians that find themselves on the ocean get caught in that filter and wind up dead.

  Measey, Drewes, and the rest are not claiming, though, that a lot of rafts carrying amphibians have reached São Tomé and Príncipe. They know that, for any particular raft one might choose as it slides off into the water from the banks of the Congo, the odds of delivering live amphibians to the Gulf of Guinea islands are almost nil. But in a high flood year (which during some periods might be every year), dozens of rafts with this remote chance of succeeding might form. Multiply that by the number of such years since the Congo and the northward-flowing ocean current have existed, and you’re talking about many thousands of rafts with amphibian passengers. It would take only a few successful ones to account for the six frog species and the cobra bobo now found on São Tomé and Príncipe.

  Bob Drewes became sufficiently enthralled with this scenario that he got an artist friend to make a painting of a natural raft drifting north in the Gulf of Guinea (see Figure 8.6). In Drewes’s mind and in the painting, the raft is more like a floating island, complete with a small forest and a meadow atop tan-colored cliffs. There’s even a stream flowing off the back end of the island. The way the artist has painted the flow of ocean water makes it look as if the raft-island has a hidden motor, but other than that, the depiction is realistic. On its current course, the raft-island seems to be headed straight for a real island on the horizon. Presumably that island is São Tomé, and the floating island is carrying a few bright yellow caecilians and generic-looking frogs, the ancestors of the modern cobra bobo and Ptychadena newtoni.

  8.6 The colonization of São Tomé and Príncipe intrigued Bob Drewes so much that he commissioned an artist friend to paint a floating island, a likely means of dispersal for amphibians and other groups. Painting by Richard E. Cook.

  THE IMPORTANCE OF THE POSSIBLE

  “We cannot prove this happened,” Bob Drewes wrote of the hypothesis that amphibians rafted from East Africa down the Congo and then up to São Tomé and Príncipe. I don’t think he meant this in the philosophical sense of nothing ever being known with absolute certainty in science. I suspect he meant it in a more practical sense along the lines of “Since we can’t build a time machine, the hypothesis will never become an accepted fact.” We may accept that some sort of oceanic dispersal is required to explain why there are amphibians on São Tomé and Príncipe, but it is a big step from there to buying the entire convoluted scenario described by Measey and his colleagues. There are other ways to explain the distributions of the Gulf of Guinea amphibians and their relatives, and we will probably never be able to rule them all out. For some scientists, this might sound like a reason not to pursue the subject. Why spend a lot of time on a question you can never answer definitively? In fact, the whole enterprise of fleshing out this dispersal story doesn’t even quite seem like “Science” with a capital “S,” at least not in the way we were taught in our high-school or college classes. Where’s the list of alternative hypotheses and the critical tests to try to falsify each one?

  To see where this and many other dispersal scenarios fit into a scientific argument, it’s worth consulting a philosopher. Not, as it turns out, a well-known philosopher of science like Karl Popper or Thomas Kuhn (and that pretty much exhausts the list of well-known philosophers of science), but a fairly obscure philosopher of history named William Dray. In particular, we’ll consult a 1957 book Dray wrote with the straightforward, academic-sounding title Laws and Explanation in History. In that book, Dray pointed out that it’s very common for an explanation of an event to consist of simply showing how the event could have occurred, without ever claiming that is how it actually did occur. That kind of explanation is useful when the event otherwise seems unbelievable; the explanation converts the listener from thinking that something could not possibly have happened to thinking that it might have happened. Dray called this kind of account a “how-possibly explanation,” as opposed to the usual “how-­actually explanation.”

  Whether one knows the term or not, how-possibly explanations are familiar to all of us in one form or another. For instance, consider the following: Witnesses say that the suspect was at a party when the murder took place. But it was a costume party and everyone was wearing a mask until the end of the evening, when identities were finally revealed, so no one can say for certain exactly when the suspect was there and when he was not. An accomplice could have been behind the mask, while the suspect was off killing the victim. That’s a standard sort of argument used by prosecutors and mystery writers to show that a suspect could have done something that at first seems to have been ruled out. It is a how-possibly argument. A bit closer to our subject, the famous “origin of life” experiments by the chemist Stanley Miller in the early 1950s, in which he mixed together molecules thought to have been present in the Earth’s early atmosphere, ran an electric current through the mixture, and ended up with organic compounds, including amino acids, is part of a how-possibly argument. Miller didn’t claim that he had recreated an actual step in the origin of life, but his experiments helped overcome disbelief about one part of that process, the natural generation of organic compounds from inorganic ones. Once we know how to recognize a how-possibly argument, we see them all over the place.

  In the case of the Gulf of Guinea amphibians, the event we want to explain is how caecilians and frogs reached those islands all the way from East Africa, a journey that on the face of it seems absurd. The explanation, the one I have just described, is essentially a series of arguments aimed at overcoming the incredulity of the audience. It sounds like the amphibians would have needed an extremely large natural raft. Well, extremely large rafts form all the time. Look at the aerial photographs of the Amazon delta. Wouldn’t a raft reaching the mouth of the Congo just get stuck in the shallows or drift off into the heart of the Atlantic? The shallows in the Congo delta are not all that extensive, and the Congo Current would push a raft north toward the Gulf of Guinea, not immediately west toward South America. How could amphibians, with their vulnerability to salt water, survive a long sea voyage? The oceanic part of the voyage might only take two weeks, and the fresh water from the Congo and the Niger during the rainy season could make the surface waters of the ocean much less salty than normal seawater.

  How-possibly scenarios like this have a long history in evolutionary biology. In fact, as a biologist/philosopher named Robert O’Hara pointed out in 1988, The Origin of Species can be read as one long how-possibly argument aimed at erasing the reader’s disbelief. To drive this point home, O’Hara makes a list of how-possibly questions and Darwin’s answers to them. For instance: “How possibly could evolution have occurred, since there is no force to drive change? Darwin removes the objection with the introduction of natural selection. How possibly could evolution have happened in so short a time? Darwin tells us that the earth is older than we thought. How possibly could evolution have taken place if we don’t see all the intermediat
e stages? Darwin tells us about extinction and the imperfection of the fossil record.”40

  The last entry in O’Hara’s list brings us back to our specific subject, ocean crossings: “How possibly could species isolated on islands be descended from other species? Darwin tells us about the powers of dispersal.” This is where Darwin brings in his seeds-in-seawater experiments, the snails on the duck’s feet, the icebergs. In so doing, he becomes the most conspicuous contributor to a long line of how-possibly arguments in the study of dispersal. For instance: Alfred Russel Wallace wrote of accounts of large natural rafts covered with vegetation drifting among the Moluccas and the Philippines; in the 1930s, a South African biologist named John Muir speculated that seeds might be carried on pieces of floating pumice; a primatologist named Anne Yoder recently suggested that lemur ancestors might have gone into a kind of hibernation to survive the trip from Africa to Madagascar; Robin Lawson and I—only dimly aware that we were following in the footsteps of many others—pointed out that the garter snakes that crossed the Sea of Cortés are especially resistant to the desiccating effects of salt water. This tradition, I think, is no fluke; it’s not a coincidence that scientists studying dispersal have come up with more than their share of how-possibly scenarios. Rather, it seems to me that advocates of dispersal, especially when invoking oceanic dispersal, are constantly faced with a very basic and widespread disbelief: How in the world do beeches, snails, lemurs, frogs, caecilians, freshwater snakes—you name your unlikely dispersing group—make ocean voyages without the aid of humans? It sounds improbable, mysterious, even miraculous. How possibly? We’ll show you how.