The creationist’s nightmare: evolution in action

Over at the Atlantic, Ed Yong shows and describes some stunning videos of “evolution in action”: in this case bacteria evolving resistance to antibiotics. It’s a clever way to visualize the accumulation of mutations over time as bacteria evolve to survive increasingly large doses of antibiotics. Beyond this demonstration, the experiment also permits serious scientific study of the nature of those mutations. Questions, for example, could involve “Do the same mutations get fixed over and over again in independent trials?”; “Does evolution ever fail to occur?” (for example, certain strains of Streptococcus bacteria in humans didn’t evolve resistance to penicillin for years, though that may now be happening);  “Are multiple mutations ever responsible for a single advance, so that there’s a waiting time for their accumulation?”, and so on.

The setup for these videos, and the conception of the experiment, was by Michael Baym at Harvard. His team built a huge petri dish, four feet by two feet, filled with agar colored black (to visualize the bacteria). At the edges of the “plate,” as shown below, there was no antibiotic in the agar. Then, as one moved toward the middle, antibiotic concentrations increased in a logarithmic manner, until in the middle there was a thousand times the amount of antibiotic that would kill the bacteria initially (the amount that would kill nearly all of them at first is the “1” stripe in the screenshot below).


Plates were then inoculated with E. coli at the two ends and allowed to adapt to the antibiotic by mutations and natural selection. They could grow toward the center only as resistance mutations accumulated. The bacteria are light colored so you can see the evolutionary wave of advance.

Here’s Ed’s description of what happens (video below) when the bacteria are challenged with ciprofloxacin—a very powerful antibiotic used for a variety of infections; I always take “cipro” with me when traveling overseas. “Real time” here means 14 days of evolution.

At the start of the video, bacteria are dropped into the edges of the dish and soon colonise the outer safe zones. Then they hit their first antibiotic wall, which halts their progress. After a few moments, bright spots appear at this frontier and start spreading outwards. These are resistant bacteria that have picked up mutations that allow them to shrug off the drug. They advance until they hit the next antibiotic zone. Another pause, until even more resistant strains evolve and invade further into the dish. By the end of the movie, even the centre-most stripe—the zone with the highest levels of killer chemicals—is colonised.

And Baym’s caption:

The MEGA-plate with a CPR gradient as in fig. S1 (0-20-200-2000-20000-2000-200-20-0). Movie was compiled from time-lapse imagery every 10 minutes for 14.2 days, and played at 30fps (18000X speed).


The second video shows adaptation to the antibiotic tremethoprim, and in this case you can see secondary mutations that speed up growth arising within one segment of the gradient. Again I quote Ed’s piece:

Resistance doesn’t come for free, and the same mutations that make bacteria invincible tend to slow their growth. You can see that in the movie below: at the 0:30 mark, the bacteria have advanced into the first antibiotic zone, but their colonies are faint and sparse.

But as the movie continues, bright spots start appearing within the faint areas. These are bacteria that have picked up “compensatory mutations”, which allow them to grow quickly and resist antibiotics. They ought to have been the fittest microbes on the plate, able to colonise new areas more effectively than their slower-growing peers. But more often than not, they became trapped. Weaker strains at the front of the expanding wave of microbes were already gobbling up all the nutrients, leaving their faster-growing peers with nowhere to grow. “You don’t have to be better than everyone else around you; you just have to be the first in a new area,” says Baym.

Here’s Baym’s caption for the video below; in this case “real time” is over about 12 days:

The MEGA-plate with an exponential trimethoprim gradient (0-3-30-300-3000-300-30-3-0 MIC). This movie was compiled from time-lapse imagery every 10 minutes for 11.7 days, and played at 30fps (18000X speed). Each second of video is approximately five hours of real time. Condensation on the lid is visible in the first several frames, and a single contaminating colony appears on the plate.

So what do we have here? As I said we have, “Evolution in Real Time”: something that creationists like Ray Comfort are always saying we don’t have. For evolution to be deemed true, say many creationists, we have to see it happen before our eyes, within a human lifetime or, preferably, within days! Yet that’s exactly what we have here, and we’ve known about this ever since antibiotics were widely dispensed after World War II. Antibiotic resistance is the paradigmatic example of evolution in real time. But we have similar real-time examples: herbicide resistance in plants, insecticide resistance in insects, and so on.

But of course, creationists don’t buy this as compelling evidence for evolution. The kind of evolution we see in the videos above, they say, is “microevolution”: one species simply changes a tiny bit to respond to a challenge. In other words, it’s evolution, but it doesn’t turn a bacterium into a dog, much less a eukaryote. What we want, say creationists, is “macroevolution in real time”: some substantial change that we see in real time—though they never define what they mean by substantial.

But the call for macroevolution in real time is impossible to meet, for the pace of such change is very slow. Nevertheless, we can actually see macroevolution over evolutionary time: big transitions in the fossil record. We have transitions between fish and amphibians, amphibians and reptiles, reptiles and mammals, reptiles and birds, terrestrial grazing mammals and whales (only about 8 million years!), and so on. No matter what you consider to be macroevolution, these are macroevolutionary changes visible through the strata. To say that they don’t count because we don’t see that change happening in a decade or so is simply bogus. Historical records of change, properly documented, are evidence every bit as valid as seeing bacteria evolve in petri dishes.

As for the ability of selection to produce big changes in short periods of time, just look at all the breeds of dogs, all descendants of a single wolf ancestor about 15,000 years ago. And if the breeds were known only as fossils, they’d be regarded not just as different species, but sometimes as different genera. Of course, creationists would respond that that’s not natural selection but “intelligent design,” since humans chose what features they wanted.  But that’s also irrelevant, for, as Darwin realized, artificial selection is a very good model of natural selection—but with humans rather than nature imposing the criteria for fitness. If artificial selection works, and works to cause big changes, then there’s no reason to say that there are some limits to evolution that allow microevolution but not macroevolution. That whole distinction between micro- and macro-, which has become a cottage industry for creationists, is specious.

h/t: Michael


  1. merilee
    Posted September 9, 2016 at 10:51 am | Permalink

    very cool demo

    • Heather Hastie
      Posted September 9, 2016 at 12:47 pm | Permalink

      My response exactly.

  2. Posted September 9, 2016 at 10:51 am | Permalink

    Every time I see something like this I want to remind everyone of Pagael and Joyce’s Darwinian Evolution on a Chip from 2008.

    Same kind of experiment, except they also sequenced the RNA strands that they were looking at. Showing exactly where the mutations occurred and then were able to examine those mutations by themsevles to see what the effects were.

    First, the experiment ran for 72 hours, which resulted in a 92-fold increase in the reaction.

    Second, one of the mutations, by itself, had negative effects, but when combined with either of two other mutations, enhanced those beyond what they would have been by themselves.

    It’s a really neat paper from 8 years ago.

  3. Todd J Morgan
    Posted September 9, 2016 at 10:54 am | Permalink

    That’s truly incredible. Thanks for sharing.

  4. Shea B
    Posted September 9, 2016 at 11:04 am | Permalink

    Remarkable stuff.

    Creationists will undoubtedly ignore everything you wrote here and go back to their old line (“But…But…that’s microevolution. Therefore god”).

    • rickflick
      Posted September 9, 2016 at 12:04 pm | Permalink

      The religious actually reason slightly differently. They think: “God – therefore…er..ah…that’s microevolution”. With God as the universal premise, anything is possible.

  5. Ian Clark
    Posted September 9, 2016 at 11:11 am | Permalink

    Spectacular experiment.

  6. Posted September 9, 2016 at 11:13 am | Permalink

    Thanks for posting this. I despise the “courtesy gesture” of acknowledging the noticeable effects of microevolution without accepting the history of Darwinian selection pressure. It’s like accepting nail growth while rejecting keratinization.

  7. J.Baldwin
    Posted September 9, 2016 at 11:22 am | Permalink

    I have a question about the description of the process. The narrator says that the bacteria hits a boundary and then develops a mutation. But doesn’t the random variation have to already be present? In other words, as the bacteria replicate, random, blind mutations occur that then happen to later provide a resistance to a novel environment that selects the pre-existing mutation? His description sounds like the novel environment causes the mutation instead of selecting the mutation. Maybe this is a quibble but I’m not a trained biological evolutionist. I just want to understand.

    • Posted September 9, 2016 at 11:35 am | Permalink

      It can either be present, or occur while the bacteria are dividing on the margins. One reason I think the mutations occur near the boundary is because if they were already there, one wouldn’t expect a “delay” while colonies sit at the boundary. It’s as if they’re biding their time until a new mutation occurs.

      • J.Baldwin
        Posted September 9, 2016 at 11:49 am | Permalink

        That makes complete sense. So is it possible that any particular organism might hit a boundary and never develop a mutation fit to the novel environment? I mean, there’s no law that says that those bacteria would certainly develop a successful mutation, right?

        • eric
          Posted September 9, 2016 at 12:01 pm | Permalink

          Yes. If you watch the video you’ll see that a few spots are responsible for the majority of the next region’s growth. The areas of growth that don’t proceed through to the next region are the sorts of strains you’re talking about.

          • Mark Sturtevant
            Posted September 9, 2016 at 12:46 pm | Permalink

            I was going to suggest this too. Also, there are was lineage that broke thru the first boundary first, and descendants of that lineage continued to lead the pack. It would be interesting to see what their particular mutation was, and whether it helped them surmount the later challenges.

      • eric
        Posted September 9, 2016 at 11:59 am | Permalink

        I think if you pay close attention, you’ll see that one ‘daughter’ strain from the right crosses each barrier relatively smoothly (i.e., with little pause). So I think what we’re seeing here is a mix of strains that randomly develop a mutation helpful for resistance before they contact the next region, and strains that develop one after.

        Excellent video. The fact that the phylogenetic tree between parent and daughter strains is *literally visible* is particularly impressive and compelling.

        • rickflick
          Posted September 9, 2016 at 12:09 pm | Permalink

          Yes, the patters are particularly fascinating. They resemble some graphic images used to illustrate various growth, spreading, or evolutionary phenomena. Something like tree branches and root systems or the movement of sand in water currents, dunes on Marse, etc.

      • steve oberski
        Posted September 9, 2016 at 12:03 pm | Permalink

        I would guess that the mutation that allows a bacteria to survive across the boundary is equally likely to occur anywhere in the petri dish, but it is most noticeable when it occurs at the boundary.

        I would also guess that the colony does not actually sit at the boundary awaiting a favourable mutation, but individual bacteria are constantly trying to move across the boundary and dying in the attempt until a favourable mutation arises.

    • Ken Pidcock
      Posted September 9, 2016 at 12:18 pm | Permalink

      Many of the mutations that allow survival at higher concentrations will make the bacteria poor competitors at lower concentrations. It is only at the boundaries that those mutants will outgrow competitors.

      I’d like to see experiments where the highly resistant mutants are put in chemostats, sans antibiotic, with the wild type. I’d bet they die off right quick.

    • gravelinspector-Aidan
      Posted September 9, 2016 at 1:27 pm | Permalink

      When the bacterial “white” (I guess they chose a strain to contrast with the agar, or the agar to contrast with their chosen strain) gets to a boundary, it doesn’t stop dividing. The colonies are continuing to grow and so to divide, and so generating new mutations.
      I have a quibble over this. If the antibiotics used work by a “catalytic” process in which the antibiotic isn’t consumed, then why would the concentration of antibiotic matter once there is enough for (in the limit) one molecule per bacterium. (More realistically perhaps one molecule per genome? Or per ribozyme?)
      OTOH, if the antibiotic is consumed by the bacterium, then simple persistence of the bacteria will eventually lower the concentration at a particular site to the point at which the bacteria can tolerate it’s presence. Then the bacterial front will advance at a rate dictated by the rate of consumption of the antibiotic by the bacteria.
      Ah … watching the videos by following the link to YouTube, I get the commentary, and on the second one at least I see the process I describe above (at the RH 10-100 boundary, a little above the half-way mark, 00:35 to about 00:40). And it is much slower than the up-gradient or cross-gradient progress of the mutant strains. So there are at least two processes going on here.
      The commentary makes the point that the agar was in two layers – the black antibiotic-dosed agar then on top of it a transparent (-ish) non-antibiotic agar in which the bacteria can live, if they can tolerate the diffused antibiotic. That diffusion is going to complicate matters.
      According to YouTube, only 180 views. People (me included) scream about having to pay for access to (generally) public-funded science. And when things like this are made available, it’s only a very few nerds who make use of the facility. I don’t know about other people, but I spend the thick end of GBP 200 a year on professional affiliations that I don’t really need, in large part because it also gives me an OpenAthens login which lets me into a lot of other sites in “the literature”.

  8. Posted September 9, 2016 at 11:53 am | Permalink


  9. Ken Kukec
    Posted September 9, 2016 at 12:05 pm | Permalink

    … it doesn’t turn a bacterium into a dog, much less a eukaryote …

    I will gladly defer to your expertise, Perfesser, but under the three-domain taxonomy, don’t even d-o-g-s qualify as Eukaryota?

    • Luis Servín
      Posted September 9, 2016 at 12:12 pm | Permalink

      Ha, ha!! Nice catch, I didn’t notice that when I read it!!

    • gravelinspector-Aidan
      Posted September 9, 2016 at 1:32 pm | Permalink

      D*gs are sub-eukaryotes. Because cats are eukaryotes.

    • Diane G.
      Posted September 12, 2016 at 2:31 am | Permalink

      I smiled at that, too. Obviously Jerry meant to use the two terms in reverse order. The subconscious can be so revealing.😉

  10. Posted September 9, 2016 at 12:44 pm | Permalink

    What is wrong with you people? Don’t you know that this video isn’t real. Its just a computer-generated film made by a member of the athiest conspiracy of scientists!!! Also, the fossills were put there by God to test our faith. Everyone knows that!!! Evolution is a hoax.

    That was an actual response I once received from a high school acquaintance of mine.

  11. Mark Sturtevant
    Posted September 9, 2016 at 12:52 pm | Permalink

    As others have pointed out, the branching pattern where some lines advance more quickly than others is very striking, and it reminds me of this famous passage:
    “As buds give rise by growth to fresh buds, and these, if vigorous, branch out and overtop on all sides many a feebler branch, so by generation I believe it has been with the great Tree of Life, which fills with its dead and broken branches the crust of the earth, and covers the surface with its ever-branching and beautiful ramifications.”
    C. Darwin

  12. E.A. Blair
    Posted September 9, 2016 at 1:04 pm | Permalink

    As long as the bacteria are so much faster at adapting than we are, we need evolutionary studies to stay ahead of them.

  13. Eddie
    Posted September 9, 2016 at 1:05 pm | Permalink

    The micro/macro evolution distinction seems a failure in basic assumptions related to cumulative changes in a system.

    Examples of cumulative change effects in systems that are not biological in nature might be helpful on this front rather than just focusing on the semantics of the word “kind.”

    • Diane G.
      Posted September 12, 2016 at 2:36 am | Permalink

      Like, say, a study of Windows from 1985 on? Or perhaps computer languages.

  14. gravelinspector-Aidan
    Posted September 9, 2016 at 1:35 pm | Permalink

    FTFAbstract :

    Analyzing mutants at and behind the propagating front, we found that evolution is not always led by the most resistant mutants; highly resistant mutants may be trapped behind more sensitive lineages.

    Or in HuffPo Clickbit speak : “Bacterial Sunday Drivers hold You Up!”

  15. Chris G
    Posted September 9, 2016 at 2:15 pm | Permalink

    Jerry – did these experiments answer the questions you pose in your first paragraph: ‘Do the same mutations get fixed over and over again in independent trials?’ and ‘Does evolution ever fail to occur?’
    i.e. were the genomes, of the bacterial-variations that evolved, analysed to determine the exact mutations that took place, and whether the same set of mutations were duplicated from both left and right in each experiment?
    At what point does a set of variations constitute a new species/macro-evolution for bacteria?
    Chris G

    • Gregory Kusnick
      Posted September 9, 2016 at 2:49 pm | Permalink

      The videos show plenty of lineages that did not make it across the barriers. So that’s one possible sense in which “evolution failed to occur” for those lineages.

      If we take the question to mean whether it’s ever the case, in replications of this experiment, that no lineages make it out of the first band, I think that’s very unlikely given the low barrier between the first and second bands.

      If that first barrier were much steeper, and went directly from zero to (say) 1000, that would be a different story, and I wouldn’t be surprised if no lineages were able to make that jump in one go.

      • Chris G
        Posted September 9, 2016 at 3:24 pm | Permalink

        Gregory – given the increments across the different ‘zones’ of concentration were logarithmic, the jump from 100 to 1000 was effectively bigger than from 0 to 1.
        And the time-scales of ‘travel’ across the boundary of each zone is uniform, which implies moving from 100 to 1000 wasn’t any more ‘difficult’ than from 0 to 1.
        Also, should we assume that a move from 100 to 1000 would necessarily require more mutations than a move from 0 to 1 i.e. that a greater change in environment must involve a greater change in genes? This is where analysis of the actual differences in genomes would be very interesting,
        Chris G

      • Sixtus
        Posted September 9, 2016 at 5:46 pm | Permalink

        Your last paragraph gives me the idea that this type of experiment (with a very high first barrier) could provide a graphic illustration as to why one is supposed to complete one’s antibiotic treatment even though you may feel better before using up all the drugs.

    • nicky
      Posted September 10, 2016 at 11:12 pm | Permalink

      Treponema pallidum (the ‘Syphilis’ bacterium) apparently always remains sensitive to penicillin, although the doses used now are higher than before, particularly in neurosyphilis. That might have more to do with penetration and pharmacokinetics than resistance.
      Will resistance (n)ever occur in Treponema? We don’t know.

      • Diane G.
        Posted September 12, 2016 at 2:39 am | Permalink

        That’s really fascinating!

  16. Bessemer Mucho
    Posted September 9, 2016 at 2:20 pm | Permalink

    Whatever mutation it was that made it to the central zone, I think I caught it last year.

  17. Todd Steinlage
    Posted September 9, 2016 at 3:59 pm | Permalink

    I wonder how many of the antibiotic resistance genes are on plasmids? In many bacteria, plasmids are (relatively) easily shared, even between species. Very cool demonstration.

  18. Kopper
    Posted September 9, 2016 at 4:20 pm | Permalink

    Th rooms where these giant plates were must have been quite stinky at the end

  19. keith cook +/-
    Posted September 9, 2016 at 4:36 pm | Permalink

    All i can say is, it didn’t grow eyes and ask the experimenter what he would like for his birthday…
    Bacteria are good at evolving but useless for social occasions.

  20. Torbjörn Larsson
    Posted September 10, 2016 at 3:56 am | Permalink

    That whole distinction between micro- and macro-, which has become a cottage industry for creationists, is specious.

    A MEGA example of seeing what you did there.

    I am reminded of Hug et al’s recent mass sequencing phylogeny, where they set up an objective threshold definition of phyla divergence. The tree recovers the two-domain tree with eukaryotes diverging from recently observed Thorarcheota next to Spang’s Lokiarchaeota. Notably eukaryotes corresponds to a single prokaryote phyla!

    I also noted this cultural difference:

    I always take “cipro” with me when traveling overseas.

    Of course Scandinavians could buy “black market” antibiotics now (though I dunno about the legality). But – speaking especially on the context of antibiotics resistance, by the way – I am pavloved to try to get a diagnosis before attempting a remedy.

    • nicky
      Posted September 10, 2016 at 11:17 pm | Permalink

      Torbjorn, self-medication with antibiotics and without diagnosis is generally not a good idea, indeed.

  21. Posted September 10, 2016 at 7:24 am | Permalink

    “You don’t have to be better than everyone else around you; you just have to be the first in a new area,” says Baym.

    And thus, Microsoft is explained.

    • Posted September 12, 2016 at 11:20 am | Permalink

      Except by most measures, Microsoft was *not* first in anything. (Not the first personal computer OS, GUI, mouse, or office suite, at least.)

  22. Pete
    Posted September 10, 2016 at 9:06 am | Permalink

    Until those bacteria can hold a banana I don’t buy it😉.

  23. Mike
    Posted September 13, 2016 at 8:28 am | Permalink

    That was beautiful, you can show a Creationist the truth, but he reserves the right to remain an ignorant Idiot.

  24. Zetopan
    Posted September 14, 2016 at 7:15 am | Permalink

    The creationist claims that macroevolution does not exist unless it can be shown in real time is equivalent to claiming that the Andes mountain range could not have formed by plate tectonics because that cannot be shown in real time. They also cannot even agree about where the borders are located that distinguish the different “kinds”. And those are just two of the myriad of problems that occur with their willful ignorance.

One Trackback/Pingback

  1. […]   houlahound, Sep 20, 2016 at 3:57 […]

Post a Comment

Required fields are marked *

%d bloggers like this: