This week’s BEACON Researchers at Work post is by University of Idaho faculty member Luke Harmon.
A day gecko, Phelsuma ornata, represents a particularly charismatic subject for the study of macroevolution.
I am a researcher who typically studies evolution over very long time scales – tens to hundreds of millions of years. For example, we’ve published papers analyzing trait evolution, speciation and extinction at scales including cichlids, day geckos, and jawed vertebrates. As such, I’ve thought a lot about my relationship to BEACON’s theme of “Evolution in Action.” Certainly we can see “evolution in action” over any time scale – but many BEACON projects focus on evolutionary change that we can observe from one generation to the next. What my work over long time scales has taught me is that it is sometimes worthwhile to “zoom out,” and to investigate the long-term consequences of processes acting over extraordinarily long time scales. One fantastic example of the value of such an approach can be seen in the Lenski lab’s long-term experimental lines of E. coli – a simple experiment which, repeated over twenty-five years, has given profound insights into evolution.
What would the BEACON logo look like if we evolved it for over a billion generations?
I believe that this “long-term” perspective applies equally well to all applications of evolution, including both biological and digital. In this essay, I will explore what I view as some of the main lessons of evolutionary studies over long time scales. As computers become faster, digital evolution will undoubtably be more akin to “macroevolution” – evolution across the entire history of the Earth – than what might happen in a few generations as a population adapts to its environment.
First, some background: evolutionary biologists sometimes use the term “microevolution” to describe evolution within a population over a few generations. This is sometimes contrasted with “macroevolution,” which can mean evolution over very long time scales, or sometimes evolution “above the level of species.” There has been some debate about the relationship between these two levels. For example, most evolutionary biologists believe that macroevolution is simply microevolution “writ large” – that is, macroevolution is what happens when you scale microevolutionary processes over extremely long time scales. By contrast, some scientists believe that there are distinct macroevolutionary processes that cannot be described in terms of the common set of microevolutionary models. The best example of this way of thinking is Stephen J. Gould, whose theory of punctuated equilibrium supposed that all sustained trait change occurred at the moment of speciation. At other times, according to Gould, lineages underwent evolutionary stasis – we might see them change a little from one generation to the next, but these changes were almost always transient and did not help explain broad-scale patterns that we see in the fossil record and across the tree of life. To Gould, a key property of punctuated equilibrium is that higher-level processes – that is, selection among units that are more inclusive than individual organisms – dominate long-term patterns.
Gould’s theories – and related ideas of species selection – are, as you might expect, controversial. It seems clear to me that punctuated equilibrium is not a general pattern – although the model is still considered by quantitative paleontologists and others. And species selection, though revitalized by new phylogenetic approaches, is still controversial. But Gould illustrates that there are some questions that can really only be addressed by studying evolution at the broadest scales. Furthermore, Gould’s observation that “stasis is data” is a key idea in the field.
I think that modern macroevolutionary studies can provide three relatively simple take-home points that are relevant when thinking about what might happen when we scale up our experiments, observations, or digital organisms over very long time scales:
1. Short-term rates do not usually scale up to longer time scales. One peculiar pattern that seems to be shared across evolutionary studies over a huge range of temporal and spatial scales is that the ‘apparent’ rate of evolutionary change depends strongly on the time scale over which those rates are measured. As first observed by Phillip Gingerich over 30 years ago, we see the fastest rates of evolution over the shortest time scales. However, this rapid change does not translate into large amounts of change over longer time scales; much of the change we see over short time scales seems to be ephemeral. This may be due to rapid reversals in selection, so that the change we see in one generation is undone in the next.
2. Life is characterized by high turnover. On average, over the history of the earth, speciation and extinction are nearly in balance. Clades have periods of growth and collapse, and taxa wax and wane; but on average, speciation and extinction are – to an approximation – equal in rate. Even more intriguingly, there is some evidence that speciation and extinction tend to change in lockstep with one another – that is, things that increase the speciation rate of a clade also tend to increase the rate of extinction; and things that prevent extinction might also slow down speciation rates (see Stanley’s interesting book, Macroevolution, for more details on this).
3. Novelty takes time. We do not yet know as much as one might hope about the evolution of novel traits – although new developments in genomics and developmental biology will likely lead the way to a deeper understanding of novelty in the future. Nonetheless, the branches of the tree of life are marked by a few key, but extremely rare, events: the origins of flight, endothermy, photosynthesis, and others. These and other rare characteristics sometimes provide the only evidence we have for deep short branches in the tree of life.
I’m not sure whether these ideas are helpful to those of you who study “evolution in action.” I don’t think it would be hard to argue that the temporal “scale” of digital evolution is constantly being compressed, and that such experiments really live in the world of “macro” – rather than “micro” – evolution.
For more information about Luke’s work, you can contact him at lukeh at uidaho dot edu.
