Friday, November 27, 2009

Revenge of the Phantom Monkeys

Over the last week or so, there has been much chuntering (or perhaps chuckling) on the VP listserve regarding a paper by Michael Heads that was published in the latest edition of Zoologica Scripta. Entitled "Evolution and biogeography of primates: a new model based on molecular phylogenetics, vicariance and plate tectonics," Heads' paper makes the remarkable claim that primates originated in the Early Jurassic, 185 MYA. That's a mere 130 MYA before the appearance of the first primates in the fossil record. Heads further concludes that the split between strepsirrhine primates (lemurs, galagos, lorises and - of course - our old friend Darwinius maxillae) and haplorhines (monkeys and apes, including us) occurred 180 MYA, that Lemurs arrived in Madagascar 160 MYA, and that New World and Old World monkeys diverged 135 MYA. What is particularly awesome about this paper is that there is no evidence whatsoever for any of this, and yet he still managed to get it published.

This isn't to say that his argument isn't worthy of consideration. It draws on two thorny problems for evolutionary biologists. This first of these is answering the question of how modern groups of animals and plants ended up with the global distribution that they have today. Back in the days before plate tectonics, biologists used to postulate the existence of vanished "land bridges" connecting continents or seemingly unlikely models of colonization such as "rafting" (animals get swept out to sea on logs or other floating debris and get washed up in novel places). Then in 1915 German geophysicist Alfred Wegener proposed that the continents themselves might drift around the surface of the planet carrying animals and plants with them and after everyone had finished laughing at him (which took around 40 years) they came to realize that this was actually true. Of course, Wegener had been dead for over 20 years at this point, having expired from a heart attack on top of a glacier in Greenland, so I guess some people may have felt a bit bad about this. But hey, that's science!

The problem is that continental drift can't account for all animal distributions. When you reconstruct the positions of the continents at critical periods in evolutionary history, it becomes apparent that sometimes animals had to cross water to get to colonize new areas. What it can do is make the gaps between continents narrower than they are today, or create island chains through volcanic and other tectonic processes, which then allow species to "hop" from island to island like stepping stones. But even so, you are still left with models that require animals to either swim or float for quite long distances and for some people that's just too unlikely to swallow.

In the case of the primates there are two major instances where this has to be invoked as a scenario to explain modern distributions - first, the colonization of Madagascar by the ancestors of today's lemurs, which are believed to have arrived from Africa somewhere between 50 and 80 MYA, and second the origins of South American monkeys, which are also thought to have evolved from African ancestors around 35 MYA. In each case, the conventional model of biogeography says that they did this by rafting across the Mozambique Channel and the Atlantic respectively, both of which were narrower at the time when these events occurred.

Phooey - or words to that effect - says Heads. The most parsimonious solution to this problem, given that Africa was at one time connected to both South America and Madagascar, is that the ancestors of these groups were once part of a widely distributed population of proto-primates that fragmented as the continents broke apart. That being the case, you can date the origins of each group of primates, plus some of their close relatives, to the radiometrically determined dates for the separation of the continents. Which all occurred during the Jurassic.

There are certainly animals that look like mammals around in the Jurassic, but none of them are particularly like the modern-day groups and many palaeontologists prefer to call them "mammaliaformes" to differentiate them from the "crown group Mammalia" as the true mammals are known. There's certainly nothing remotely resembling a primate; the earliest known eutherian mammal, Eomania, is 125 million years old. That makes it about the same age as the earliest known ancestor of marsupials, Sinodelphys, and 2 million years older than the oldest monotreme, Teinolophos.

This brings us to the second thorny problem, which is what to do about dating evolutionary events that must have happened (e.g. the origin of primates) but for which we have no direct fossil evidence. Finding the oldest fossils in different lineages is helpful because it enables us to fix the latest date for the common ancestor of those groups. So if Eomania and Sinodelphys are both 125 million years old, and Teinolophus is 123 million years old, then the common ancestor of modern mammals (monotremes + marsupials + eutherians, incl. primates) cannot be less than 125 million years old. The problem is that we can only use the fossil record to fix the most recent date. Absence of evidence, as we all know, is not evidence of absence and given the notoriously patchy nature of the fossil record there may be earlier members of these groups that are as yet undiscovered.

As ever, molecular biologists think they have an answer to this. Back in the 1960s (which is like the stone age of molecular studies) they made the observation that the molecular sequence of a variety of biological molecules, including proteins and DNA, changes over time. It follows, they hypothesized, that the magnitude of the difference between the same molecule in two different groups of organisms is a reflection of the time since the two lineages diverged. Provided that this rate remains constant, you can calibrate your "molecular clock" to figure out how long ago the common ancestor of the two groups. Once again hard science triumphs over the woolly uncertainties of the fossil record. Hurrah!

Of course, it's not as simple as that because the clock does not "tick" at the same speed in different lineages. Albatrosses have molecular clocks that run at half the speed of other birds. Small mammals accumulate molecular changes at 8 times the speed of turtles. Over time and much study it's become apparent that there are several factors affecting the rate of molecular change, including population size, the time taken to reach reproductive maturity, the intensity of natural selection, and physiological differences between species. This means that you have to apply a series of statistical and computation fudge factors to your analysis to compensate for these confounding variables. In short, the molecular clock can provide supporting evidence for a hypothesis, but is not direct evidence. This is a pity for Heads, because the molecular clock for primates does suggest an earlier origin for primates than the fossil record - but only 80-116 MYA, which is still a good way short of his 180 MYA model. Having said so, there's no more evidence for 80 million year old primates than for 180 million year old ones.

Heads' Jurassic primates are a classic example of a "ghost lineage." Ghost lineages are phylogenetic branches that are inferred to exist, but for which there is no direct (i.e. fossil) evidence. Ghost lineages are a plague for those researchers interested in the early history of the modern placental mammal groups because molecular dating suggests that they originated back in the Cretaceous, but so far no-one has found a fossil member of one of these groups that predates the K/T boundary 65 MYA. Open-minded paleontologists (and there are a few of these rare creatures around) are prepared to at least consider the possibility that there might be some late Cretaceous primates kicking around. But no-one is going to swallow the idea of an unrecorded evolutionary history for primates that is almost three times longer than their known fossil record.

Why? Well, the fossil record for Mesozoic mammals isn't great, but it's still good and we know from the fossils that there were other other animals of similar size and ecologies (e.g non-placental eutherians) known from the "missing" time period. That means that you have to start coming up with increasinglt tenuous arguments to explain why the animals in question (e.g., Jurassic primates) are completely missing. Plus, if you accept that there were primates in the Jurassic, then you also have to conclude that the diversification of almost all the groups of placental mammals took place even earlier than this. Not being able to find a fossil of a Jurassic primate is bad enough; not finding evidence of *any* modern group if their ancestors were all present makes this hypothesis untenable.

No comments:

Post a Comment