A complex organism is better than a simple one, so natural selection favors complexity--or so it has been assumed. Is the assumption true?
Everybody knows that organisms get better as they evolve. They get more advanced, more modern, and less primitive. And everybody knows, according to Dan McShea (who has written a paper called Complexity and Evolution: What Everybody Knows), that organisms get more complex as they evolve. From the first cell that coalesced in the primordial soup to the magnificent intricacies of Homo sapiens, the evolution of life--as everyone knows--has been one long drive toward greater complexity. The only trouble with what everyone knows, says McShea, an evolutionary biologist at the University of Michigan, is that there is no evidence it’s true.
At one level, of course, it must be true: we really are more complex than that first cell, and we’re not alone. Anyone can see that the first organism on this planet was simpler than a modern reptile, or beetle, or whatever, says McShea. But that proves very little: the first organism was almost certainly the simplest organism--which means that in terms of complexity, there was initially no place for life to go but up. The interesting question is what happened later. Did natural selection drive organisms onward and steadily upward, toward ever greater complexity, because being more complex improved their chances of survival? Researchers have always assumed the answer was yes.
But lately McShea and a few other researchers have been trying to test that unshakable assumption with real data. And the data tell a different story: evolution does not necessarily drive life toward greater complexity. As organisms evolve, they are just as likely to shed complexity as to gain it--at least insofar as biologists can measure it.
Measuring complexity is not easy because defining it isn’t, particularly in the case of a whole organism. Finding a quantitative way to compare the complexity of radically different organisms is even harder. (Is a whale more complex than a peach tree or a praying mantis?) I don’t know how to measure the complexity of whole organisms--I don’t even know what that means, McShea says. And no one else does either. To make the problem tractable, McShea decided to compare only like organisms with like and to look not at the whole organism but at a particular part.
He chose vertebrates, in particular the vertebral column, because they were convenient. (For one thing, there are lots of vertebral columns lying around natural history museums, waiting to be measured.) His premise was straightforward: a vertebral column consisting of, say, 20 vertebrae, all of them identical, is a simple vertebral column. If all the individual vertebrae are different, however, the column is complex. So to determine the complexity of a vertebral column, McShea measured how different the individual vertebrae were from one another along six separate dimensions-- length, thickness at the thickest point, and so on.
Then he started comparing the vertebral columns of ancestral animals with those of their descendants--relatives separated by about 30 million years of evolution--in each of five groups: squirrels, ruminants (hoofed animals such as cows), camels, whales, and pangolins (anteaterlike mammals with horny scales). If evolution drives organisms toward greater complexity, McShea reasoned, such comparisons should reveal the trend clearly. He got a clear result, all right, but not a trend: in fact, he found the complete absence of a trend in complexity.
In most of the comparisons, there was no significant change in complexity in either direction, says McShea. And the few cases in which complexity seemed to increase from ancestor to descendant were offset by complexity decreases in other pairs. The bottom line, says McShea, is that this showed no preferred tendency for complexity to increase. Increases and decreases tend to happen about as often.
As it happens, two other researchers have recently arrived at essentially the same bottom line while studying completely different organisms. Paleobiologist George Boyajian of the University of Pennsylvania and geologist Tim Lutz of West Chester University looked at ammonoids, free-swimming spiral-shelled mollusks that existed for 330 million years before going extinct with the dinosaurs some 65 million years ago. Like its surviving cousin the nautilus, the ammonoid grew by adding chambers one after another to the open end of its shell, spiraling outward like a French horn. The chambers were separated by walls of calcium carbonate called septa.
Ammonoid septa have been a textbook example of how complexity supposedly increases with evolution. The septa rarely grew straight across the inside of the shell. Instead they branched and meandered like delicate ferns, along paths that seemed to grow more intricate over the course of evolution. The outlines of the septa can be seen today in fossil ammonoid shells, preserved as a sort of tracing where the septum was sutured to the inside of the shell.
To measure the complexity of those sutures, Boyajian and Lutz took advantage of a ready-made technique: the mathematics of fractals. The fractal dimension of a line is a measure of the degree to which it is jagged or convoluted--and in that sense complex--rather than straight. Boyajian and Lutz determined the fractal dimension of 615 ammonoid sutures, one from each of 615 different ammonoid genera. Their measurements showed that the most complex sutures were indeed to be found among the later ammonoids, as the textbooks claimed.
But when the two researchers followed McShea’s example and looked at ancestor-descendant pairs, the result changed. In those pairs, says Boyajian, there’s an equal chance of the ancestor being more complex or less complex than the descendant. In other words, we don’t see any direction to the change in complexity.
Then why are the most complex ammonoid sutures also the later ones? The two results, says Boyajian, are not contradictory. Even if the evolution of ammonoid sutures is random, as the ancestor-descendant comparisons suggest-- meaning that at any given time an ammonoid is just as likely to become simpler as it is to become more complex--the maximum complexity to be found among all ammonoids will increase over time. Some ammonoids that are already complex will, by pure chance, become yet more complex even while others are becoming simpler. But the real issue is not whether complexity can increase by pure chance during evolution; it is whether natural selection is driving organisms toward greater complexity. In other words, the issue is whether a complex animal is a more successful animal.
Boyajian and Lutz tested that issue directly. If the hypotheses out there were right, there should be a correlation between how long something lived and its complexity, Boyajian says. In ammonoids a more complex septum might improve the animal’s survival odds by strengthening the shell (just as corrugations strengthen cardboard). If so, then all other things being equal, a more complex ammonoid ought to have survived longer in the fossil record than a simple one.
Yet when Boyajian and Lutz examined how long each of their 615 ammonoids had lasted, they found no benefit in complexity. Complex or simple, very few ammonoid genera lasted more than 15 million years, says Boyajian. And those that did weren’t necessarily the most complex ones. So complexity didn’t help them, but it didn’t hurt them either. The complexity of ammonoid sutures, it seems, is the result of random variation rather than natural selection.
One might object that the complexity of an ammonoid suture--or of a vertebral column--does not necessarily reflect the complexity of the organism as a whole, or the type of complexity that is most relevant to survival. Surely the tremendous but difficult-to-measure complexity of, say, the human brain has evolved because it conveys some survival benefit that a simpler organ could not?
I’m suspicious of the tendency to assume that more complex is better, says McShea. More complex can also be worse: more parts mean more things can go wrong; the organism may be harder to put together, harder to keep together. You need to test things like this.