For most of my life, my understanding of the debate over evolution involved evolutionists (who argued that random, purposeless evolution definitely happened) against creationists (who argued that random, purposeless evolution definitely didn’t happen). All seemed to accept the premise that evolution, guided by random mutation and natural selection, was random and purposeless, with no determined or inevitable outcomes (and thus no sign of divine involvement, either), with the only dogmatic disagreements being about whether or not that evolution actually happened.
Recently, however, I keep bumping into what is apparently a growing paradigm of people that believe evolution happened, but that it was not so random and purposeless! From Michael Behe arguing that the crucial mutations for the evolution of life must have been “non-random,” (Edge of Evolution review) to Perry Marshall highlighting James Shapiro’s ideas (Evolution 2.0 review) that random mutation is not actually the main driver of evolution but rather a host of other more complicated cellular abilities and more active (almost self-driven) responses to environmental challenges. For me, this all cuts across the old debate, with fascinating implications for the anthropic principle and the old assumptions about the theological implications of evolutionary history.
Like St. George Jackson Mivart, I’ve always been fascinated by convergent evolution – the idea that various features not only managed to evolve at all, but apparently evolved multiple times (due to the feature not fitting neatly into the best-possible-fit of hierarchical nesting boxes of common ancestry). From a creationist perspective, it seemed to be a potential hole in the theory, and from an Evolution-2.0 perspective, it seemed to be evidence that evolutionary development was not completely random but had some sort of predictability or determinism in its outcomes. I heard about a new book that drew on the new flood of genomic data to show how convergent evolution is far more common than anybody ever thought, and how evolution can be far more rapid than anybody ever thought, and how all these data and experiments are calling into question assumptions about the unpredictability and inevitability of evolution. This sounded to me like more of that new paradigm-busting kind of evolution. And unlike the creationists or random outlier scientists who have always been claiming that the old view of evolution was a crisis about to collapse but that their maverick ideas are ignored and shut out of the ivory towers, this was coming from a Harvard professor citing the cutting-edge work of multi-published academic colleagues – about as central to ivory-tower-world as you can get. I knew I had to read Jonathan Losos’s Improbable Destinies and see what it had to say.
Whirlwind Tour of Convergence
The opening chapters lay out the competing scientific “debate between contingency and determinism.” Stephen Jay Gould represents the old “contingent” view that evolution is slow, random, and unpredictable. It has no foresight or planning or purpose. If you changed the slightest bit of history and “replayed the tape,” there might have been a completely different chain of outcomes and humans might not even be here. The universe guarantees nothing. Simon Conway Morris represents the new “deterministic” view that evolution is repeatable and predictable, consistently managing to derive similar outcomes from different starting points. And it can be fast too; Losos says “the reality of rapid evolution” is that “evolution can rip along at light speed” when conditions change, and that “life repeats itself… evolving similar adaptations in response to similar environmental circumstances.”
There has been an exponential rise in genome sequencing in recent years, and we’re finding all kinds of animals that share features that apparently evolved separately because their DNA is too different for them both to have inherited those features (“as new data from molecular biology floods in… time and time again we’ve been misled”). The book takes us on a whirlwind tour of examples of such astonishing convergence:
- Australian marsupials have “convergent placental counterparts,” including a sugar glider instead of a flying squirrel, a marsupial mole instead of a mole, and a wombat instead of a groundhog, all with similar appearances and filling similar niches but evolving completely independently
- Losos had assumed “porcupines were one happy evolutionary family” but “learned I had it all wrong. New and Old World porcupines do not share a common evolutionary heritage… The two lineages have independently evolved their quills from different, unquilled rodent species. They are the result of convergent evolution.”
- “the traits that define the [beaked sea snake] species, not only its beak, coloration, and general appearance, but also its nasty disposition, have evolved convergently, so much so that distant relatives on opposite sides of the Indian Ocean were considered to be members of the same species” [until their genomes were sequenced]
- “many types of lizards have independently evolved flaps of skin under their necks that can be pulled out quickly… to signal…”
- The “mantidfly… has nearly identical forearms for capturing prey… long neck and bulging eyes are so similar that its front half is a virtual mantis carbon copy, even though the two insects are separated by hundreds of millions of years of insect evolution”
- “Despite their phylogenetic distance, the social structure of ants and termites is remarkably similar,” including “construction of underground fungus gardens” which include “removing waste products, controlling pests” and using “antibiotics grown from bacteria.. on their body or in their guts”
- “the lake stickleback populations… convergently lost most of their body armor and their spines shrank”
Losos especially highlights the repeated “adaptive radiation” of single populations diverging into the same niche-filling varieties in different locations, especially involving islands:
- the birds in Galapagos and Australia “radiating” into finches, wrens, blackbirds, warblers, robins, etc, not descended but “convergent” with “Northern Hemisphere families”
- “just like Anolis lizards and Mandarina snails… anatomically and ecologically different bats living in the same region were more closely related to each other” than “similar species in other regions”
However, as we’ve begun exploring the genomes of many of these convergences, we are finding that the features are not identical at the molecular level. If DNA is like a dictionary, there are multiple ways to spell many of the same words, or tell the same stories:
- On humans and milk, “different mutations – each with the same effect of keeping the lactase gene switched on – evolved in the different populations”
- “caffeine most likely evolved independently in the three types of plants” but “the NMTs modified in coffee were different from the ones modified in tea and cacao”
Thus, while all this convergence may have been unexpected by evolutionary thinking, the details and patterns of the convergence seems to be explainable by it. After describing the truly wonderful ability of anoles and geckoes to climb vertical surfaces with sticky toepads that have “millions of microscopic filaments called setae” which literally have “free electrons” that “can bond with electrons on the surface of… another object”:
- “the best examples of repeated convergence are among closely related species… Sticky toepads have evolved eleven times in geckoes, and only two other times among the more than six thousand species of lizards”
If evolution is so repeatable, so often finding the same great solutions to the same problems, so often successfully filling environmental niches with the same kinds of creatures, does that make evolutionary progress inevitable? Losos highlights Dale Russell’s arguments that, even without the infamous asteroid giving rise to mammals, selection for larger reptilian brains could have naturally led to humanoid-looking reptiles with human-level intelligence. If the fine-tuning of the universe means that “it almost seems as if the Universe must in some sense have known that we were coming,” as Freeman Dyson said, does convergence mean evolution seems to have known we were coming, too?
Well, not quite. Losos says that to learn more amazing examples of convergence you should read Simon Conway Morris’s Life’s Solution and his newer The Runes of Evolution, along with George McGhee’s Convergent Evolution. For the rest of his book, Losos dives deeper into specific examples, including much of his own work – but unfortunately the details, while fascinating in their own rights, are not quite as exciting as I had hoped – with Losos eventually throwing some cold water on the extent of convergence as well.
Lizards and Guppies and Deer Mice, Oh My
Many Caribbean islands have varieties of anole lizards, such as one species with legs and body optimized for living on the ground, one species optimized for climbing narrow twigs on low foliage, and one species optimized for living on the tops of the trees. Each island has its own unique species, but there are corresponding species on other islands with similar-looking creatures filling the same niches. Surprisingly, genomic sequencing shows that the anoles on a given island are more closely related to each other than to their corresponding niche species on the other islands, meaning that one lizard species came to each island and just happened to diverge into evolving the same features to fill the same niches on each island! He then describes experiments revealing astonishing levels of changes with these features happening within several years! Losos tells similar tales about colorful and non-colorful guppies in Caribbean pools predictably responding to the introduction or removal of predators, and he reports on other experiments as well, including a giant experiment with deer mouse in the Midwest.
This is all pretty cool, but I couldn’t help thinking that this is all what creationists would definitely call micro-evolution, and all of these examples would fit right into their baraminology of diversity within created kinds. Furthermore, the information from the genome sequencing hasn’t quite reached the potential for the really interesting stuff – like being able to tell how many mutations it actually takes to evolve different degrees of change, and how random those mutations really are. We have enough data to tell that different populations apparently convergently evolved the same features and varieties, but when it comes to identifying specific changes, especially the ones under experiments, everything seemed to be just on the cusp of identifying how many mutations they took, or which mutations were involved, like we’re almost there but the book was written just a few years too early.
Evolving E. Coli Experiments
In the next section, Losos dives into the one area where we do have that kind of data about mutations, detailing the “Long-Term Evolution Experiment” (LTEE) on E. coli, which after a few decades now involves multiple pathways of tens of thousands of generations, with old generations frozen at regular intervals to allow comparisons and repeated tests. In addition to the general improvements in the bacteria’s ability to grow and reproduce, researchers have seen the evolution of a much-touted actual new feature: the ability to use citrate instead of glucose for energy in high-oxygen environments.
Due to modern sequencing technology, we’ve retraced and identified the exact mutation involved – and for a creationist it’s actually not that impressive. The bacteria already had the ability to process citrate in low-oxygen environments, so the mutation was simply a single copying error that turned on that existing feature in high-oxygen environments, where it was found to be advantageous. The most impressive detail was that there did seem to be a couple of “potentiating” mutations that had to take place first for this final mutation to be effective, so this is arguably a bona fide example of a successful mutation that required multiple steps.
But it’s also a bit of a letdown for this to be the most exciting thing that’s been discovered after tens of thousands of generations! It’s one thing to learn how many mutations it took to turn on a feature that was already there in a new condition. I want to know how many mutations it took to build that feature in the first place! I want know how many mutations it took to evolve guppy coloration or lizard leg variation, yes, but I really want to know how many mutations it took to evolve those amazing setae vertical grips! I expect these kinds of details may have fascinating implications for whether or not evolution happened, and if so how it did so. But the knowledge is just not quite there.
As the book draws to a close, Losos delves a bit into antibiotic resistance, hoping that our knowledge of convergence will help us better fight diseases by anticipating similar evolutionary responses in a variety of species. (This made me think about Edge of Evolution, wondering how many of those resistances reflect the strong improvements of an “arms race” and how many reflect the limited trade-offs of “trench warfare”.)
As for the broader implications, Losos circles back to his opening tour, arguing that Gould’s views on contingency were misunderstood, that nature is not really inevitable enough to produce humanoid reptiles, and that convergence is not quite as repeatable or inevitable as some seem to think. Nor does he suggest that any “natural genetic engineering” tricks supersede the good old “natural selection acting on random mutation” to produce the convergences we do see (Among other things, Perry Marshall would also complain about Losos perpetuating the notion that the human eye is backwards and thus inferior to the octopus). Ultimately, Improbable Destinies caught me up-to-date on the fascinating explosion in the science of genome sequencing and evolution experiments, and all the things we’re learning from it, but it mostly left me eagerly anticipating the next levels of discovery that could truly help answer the fascinating questions that those discoveries are pointing towards.