I think about my vulnerability, just a hundred yards out from the shore of Pyramid Lake. If I had been stupid enough to swim a few hundred yards farther, I might have drowned right there, with the devil-eyed grebes drifting idly by. Take that vulnerability and extrapolate, and you see how unlikely it must be for a person to cross a large stretch of ocean without some kind of watercraft. How many times has someone made a long ocean journey—from California to Hawaii, say, or between Africa and the West Indies—without a boat? As far as anybody knows, never. Natural selection hasn’t molded us to survive for a long time in water, or even on a natural raft. It wasn’t just paranoia that made me think, as the waves pushed against me, “I better get back now.”
We have the sense, correct I think, that our primate relatives are about as bad as we are at surviving an ocean journey. The history of primate dispersal generally bears that out. For instance, no primate is known to have naturally colonized islands extremely far from other landmasses; the distance record was apparently set by lemur ancestors some 50 million years ago, when they reached Madagascar from Africa, a distance of perhaps three or four hundred miles, or possibly by monkeys that dispersed to the Greater Antilles from South America 20 million years ago or so. Macaques seem to have a special propensity for overwater dispersal, having colonized several Southeast Asian islands, such as Sulawesi and the Nicobar Islands, that are separated from other land by deep water. However, none of these macaque journeys were demonstrably more than about a hundred miles. Similarly, Homo floresiensis, the so-called hobbit whose fossil remains were discovered a decade ago on the Indonesian island of Flores, probably dispersed without boats between islands, but again the distances were short, on the order of tens of miles, not many hundreds or more.
All of this—that is, our own experiences and common sense, and the long, generally non-ocean-voyaging history of primates—contrives to make my next statement sound ridiculous: monkeys crossed the Atlantic. Think about it: I’m in danger of drowning after swimming a couple hundred yards in a slightly choppy lake, but monkeys somehow crossed the entire Atlantic Ocean. I have to admit that this case gives me pause. When I think about it, I hear Gary Nelson mocking the dispersalists for invoking the mysterious and the miraculous. I’m not one of those scientists who thinks that everything he has ever written or said is right. I’m by nature a doubter, not only of other people, but of myself, and I wonder if somehow this monkey example will show that the new dispersalism I’ve been pushing is a house of cards. I pause, then I go on to the evidence.
We can start, as we often do, with an evolutionary tree. This time, it’s the tree of primates, so most of the groups will probably sound familiar. One part of the tree, a part that won’t concern us much, contains lesser known primates, things like lemurs, lorises, and galagos. The part we’re especially interested in is essentially the monkey branch, which includes not only monkeys in the usual, vernacular sense, but also apes. A lot of scientists have spent a lot of time working on this primate tree because, of course, it is our tree. As a result, it is now a very reliable one, with much supporting evidence from anatomy and, especially, from many different gene sequences; the upshot is that the ideas about primate relationships aren’t likely to change much as new evidence accumulates.
For our purposes, a key is that one branching point in the tree leads to New World monkeys on one side and Old World monkeys (including apes) on the other. The New World monkeys include, for example, spindly spider monkeys; big, grunting howlers; and acrobatic little marmosets. It’s usually easy to recognize a New World monkey, even if you see one in an Old World zoo, because they have flat noses with nostrils that point to the sides (the scientific name for the group is Platyrrhini, which means “flat nose”), and many of the larger ones use their tails as prehensile “fifth limbs” as they make their way through the trees. The Old World monkeys and apes have a narrower snout with nostrils that point down (the name for this group is Catarrhini, which means “downward-pointing nose”). Some are short-tailed or tailless, and even those that have a long tail can’t grab hold of things with it. While all platyrrhines are highly arboreal, some catarrhines, such as baboons, chimps, gorillas, and, of course, humans, spend much of their time on the ground.
The sister-group relationship of platyrrhines and catarrhines is so well supported by a combination of anatomical and DNA sequence data that no serious student of primates doubts it. And that branching point in the tree implies a biogeographic conundrum. The platyrrhine branch leads to species only found in the New World tropics and subtropics, while the catarrhine branch leads to species only found in the Old World (with the single exception of humans). It’s another example of a disjunct distribution on a global scale, and it looks like a pattern we can explain by Gondwanan vicariance. Simply imagine that monkeys once lived in the western part of Gondwana, and that the New World and Old World lineages became isolated and went their separate ways, evolutionarily speaking, with the opening of the Atlantic.
Gondwanan vicariance—the process, if not the phrase—is what laypeople have come up with as an explanation when I’ve mentioned this monkey example. However, unlike many of the other cases I’ve described, almost no biogeographers seem to believe this scenario for monkeys. The problem with the vicariance explanation, as we have seen so often before, is that the timing is all wrong. If the opening of the South Atlantic caused the separation between platyrrhines and catarrhines, then that split in the evolutionary tree should have occurred on the order of 100 million years ago. To put this in some perspective, such an old date would imply that the New World and Old World monkey lineages, which we know are not early branches in the primate tree, are actually about 50 million years older than the earliest known primate fossils of any kind. In fact, these monkey lineages would have to be some 35 million years older than the first known fossils of any placental mammal. I know of only one biologist who is seriously arguing that the platyrrhine and catarrhine lineages are that old and were therefore subject to the opening of the Atlantic, and that one biologist is Michael Heads, the panbiogeographer who doesn’t believe in long-distance dispersal. Tangentially, it’s a bit odd that other diehard vicariance biogeographers, notorious for their disdain for almost any evidence about the ages of lineages, aren’t arguing for the Gondwanan breakup explanation in this case. Maybe this is because the timing of primate origin and diversification has been so thoroughly scrutinized that anyone arguing for the Gondwanan vicariance explanation would be immediately ridiculed.
9.4 Part of the primate evolutionary tree. All of the taxa are extinct except those marked as “Living.” Note that all of the close relatives of New World monkeys (Platyrrhini), living and extinct, are from the Old World, indicating dispersal from the Old World to the New World along the branch identified with an arrow. Dolichocebus and Branisella are fossil New World monkeys. The evolutionary tree is redrawn and modified from Ni et al. (2013). Some groups have been left out to simplify the diagram, but this exclusion does not change the interpretation.
All of the evidence about timing (described below) indicates that the platyrrhine-catarrhine split occurred when South America was either an island continent or was connected only to Antarctica/Australia (a connection that isn’t relevant to the origin of New World monkeys). Furthermore, although there are quite a few fossils in the Old World—in Africa and Asia—that might be close to the ancestor of all monkeys, there are no such fossils in South America. In the New World, monkeys appear in the fossil record as if out of thin air about 26 million years ago, which is after the platyrrhine-catarrhine split by anyone’s estimate. What this suggests is that monkeys originated in the Old World and dispersed to South America (see Figure 9.4). They had to cross water to get there, but when did they do it, and what water did they have to cross? As is so often the case, the answer to “when” is crucial to the inference of “what.”
The dating of branching points in the primate evolutionary tree—the constructio
n of a primate timetree—has received a lot of attention among evolutionary biologists. In fact, the age of divergence between New World monkeys and Old World monkeys has been estimated from molecular data in at least twenty separate studies. As I sifted through this literature, I sometimes felt like I was wading through deep mud and that there would be only confusion at the end of it. The studies estimated the age of that split at anywhere from 31 million to 70 million years, which is a very wide range of ages. In fact, it’s the kind of disappointingly wide range that might make one think that Gary Nelson was right to mock the “molecular dating game.”
Among these investigations of divergence dates, the study that especially stood out was the one that gave the age of the platyrrhine-catarrhine split as 70 million years old. That study was by a Swedish evolutionary biologist named Ulfur Arnason and several of his colleagues and used a fairly large set of data, complete sequences of eleven mitochondrial genes. It also used what, at first glance, seem like reasonable fossil calibrations for the timetree. For our purposes, the important thing is that 70 million years ago, the Atlantic was very narrow and possibly contained a large island spanning most of its breadth. So, if Arnason et al.’s age estimate is correct, and if that island really existed, monkeys could have crossed the ocean by making just two relatively short overwater journeys—one from Africa to the island, and one from the island to South America. This island-as-stepping-stone scenario would strain almost no one’s beliefs about the overwater dispersal abilities of monkeys.
The 70-million-year age is probably way off, though. In particular, Arnason et al. did two things that, in combination, probably made their age estimate far too old. First, they used only mitochondrial genes, which are notorious for evolving in a non-clock-like manner, shifting their rate of evolution strongly and frequently. They could have gotten around this problem, at least to some extent, by using several calibrations within primates; in other words, the analysis could have been done so that divergence dates were estimated using primate rates of genetic change. But Arnason et al. had a deep distrust of any primate fossil calibrations and didn’t use them, perhaps mistakenly thinking that, because the primate fossil record is generally not very good, all parts of that record must be unreliable. Instead, they calibrated their primate tree using three nonprimate branching points—whales and even-toed ungulates; horses and rhinos; and toothed and baleen whales. Those calibrations indicated slower rates of genetic change than are realistic for primates, and thus pushed branching points within the primate tree too far into the past. It was as if you estimated how long a trip would take on Amtrak when you were actually traveling on a Japanese bullet train.
If we eliminate this Arnason study along with others he did that had similar problems, the range of estimates for the platyrrhine-catarrhine split becomes 31 to 58 million years ago. Furthermore, the best studies—the ones that use large amounts of genetic data (including nuclear genes), many groups of primates, and reasonable methods to account for non-clock-like evolution—all yield estimates of 51 million years or younger. Although 31 to 51 million years ago is still a 20-million-year range, it’s a lot better than 31 to 70 million years.42 The single estimate I consider the best currently available, because of the amount of data used and the sophistication of the analyses, gives an age right in the middle of the range, 41 million years ago (with a 95 percent confidence interval of 33 to 50 million years).
Armed with these reasonable dates from molecular analyses, we can now infer the window of time in which monkeys must have dispersed to the New World (see Figure 9.5). We know that there were monkeys in the Americas by about 26 millon years ago, because this is the age of the oldest known fossil of a monkey in the New World (from the Late Oligocene of Bolivia); this sets the younger boundary of the time window. The older boundary is set by the oldest reasonable estimate for the platyrrhine-catarrhine split, 51 million years ago. Since fossils indicate that this branching event occurred in the Old World (in Africa, or possibly Asia), the dispersal to South America must have happened at some later time.
9.5 A primate timetree indicating the probable period for the colonization of the New World by monkeys (shaded bar). The oldest possible age of colonization is set by the oldest reasonable estimate for the separation between the New World and Old World monkey lineages. The youngest possible age is set by the oldest fossils of monkeys in the New World. Timetree redrawn and modified from Springer et al. (2012). Photos, top to bottom: Clément Bardot (ring-tailed lemur), Sakurai Midori (undescribed tarsier species), Bjorn Christian Torrissen (golden-headed lion tamarin), Graham Racher (chacma baboon).
The upper limit of this time window, 51 million years ago, clearly rules out Gondwanan vicariance as an explanation, a surprise to almost no one. It also rules out Arnason’s scenario of short oceanic journeys using a giant Late Cretaceous island as a stepping stone. However, on its own, the time window does not force us to accept the scenario of monkeys crossing a wide Atlantic. There is one other hypothesis to deal with, one that’s been kicking around for decades, namely, that monkeys reached the New World tropics by crossing from Asia over the Bering Land Bridge and then moving down through North America. This “via North America” hypothesis requires a journey across the Caribbean to South America, but the distance of the required ocean journey is unclear because of uncertainties about the geologic history of that region.
That uncertainty, however, is a moot point. The bigger problem with this hypothesis is that there are absolutely no fossils to support it. For a good part of that window of time from 26 to 51 million years ago, North America was a warmer and more heavily forested place than it is today, and primates are fairly common in the continent’s fossil record for much of that period, especially in the Eocene. However, they’re not the right primates. There are tarsier-like and lemur-like species of various sorts, clearly outside of the monkey clade, but there are no monkeys of any kind. So, to accept this hypothesis, we have to imagine that monkeys passed through North America but, unlike many other primates of that time, inexplicably left no trace. We know that the fossil record is incomplete, but that it should be incomplete in such a peculiar way—as if monkey fossils were selectively plucked out by some divine hand—seems extremely unlikely. This situation stands in great contrast with the fossil records of Africa and southern Asia, which contain many early monkeys or near-monkeys, some of them potentially quite close to the ancestor of the New World monkeys. Basically, the fossils strongly indicate a dramatic leap from the Old World to South America rather than the long, slow route through North America.
This rather involved consideration of the molecular and fossil evidence leaves us with a seemingly mysterious, miraculous explanation: monkeys crossed the Atlantic Ocean to colonize the New World. Here, as in many other cases in this book, we’ve come to a conclusion by refuting other reasonable explanations. In particular, molecular dating rules out vicariance via the opening of the Atlantic (or any other Mesozoic event), and the fossil record argues strongly against colonization through North America, pointing us to the path across the Atlantic.43 In reaching that point, we are following the philosopher Karl Popper, who identified refutation (or falsification) as the essence of science, or maybe Sherlock Holmes, who famously said, “Once you eliminate the impossible, whatever remains, no matter how improbable, must be the truth.”
NARROWING THE GAP
Old World and New World monkeys share several obvious features that indicate they are “sister groups,” for instance, the fusion of the two halves of the lower jaw where they meet in front, and the formation of a bony cup surrounding the eye. (In cladistic parlance, these are shared derived traits, or synapomorphies.) Nonetheless, for much of the twentieth century, many primatologists believed that these two groups were not closely related, but instead had converged on the same anatomical features from different starting points. That belief was motivated by biogeography: if Old World and New World monkeys were in fact closest relative
s, their distributions seemed to require colonization across the Atlantic, and that simply could not be. Better to interpret their several shared anatomical traits as an odd coincidence than to have to invoke such an improbable dispersal event.
The problem those scientists had with the Atlantic-crossing hypothesis is one that still seems troublesome, even with all the DNA and fossil evidence in hand. The problem, in a nutshell, is distance. It’s fairly easy to believe that macaques rafted from Borneo to Sulawesi, or that “hobbits” crossed the narrow channels separating some of the Lesser Sunda Islands. However, the Atlantic Ocean is almost 1,800 miles wide at its narrowest point, and it’s hard to imagine how monkeys could survive such a voyage. Even on a large “floating island,” wouldn’t they die of thirst or starvation long before reaching South America? Despite the evidence for this journey, shouldn’t we be trying to come up with some other, better route to the Americas?
One thing to remember is that the Atlantic, which formed as magma pushed up and spread east and west from the Mid-Atlantic Ridge, is continually getting wider. The rate of spreading is on the order of the growth rate of a fingernail, which may not sound very fast, but obviously has added up through deep time. Rewind 40 million years, to the time of the presumed simian voyage, and the ocean would be considerably narrower than it is today. In fact, at that time, the Atlantic probably was only about 900 miles wide, still a seemingly mind-boggling journey for a rafting monkey, but not as improbable as 1,800 miles.
The Monkey's Voyage Page 25