Culturing Science – biology as relevant to us earthly beings

When adaptation doesn’t happen

“Evolutionary biology has been enriched by considering not only how adaptation happens, but also why it often does not happen,
or at least does not happen as we might naively expect.”

Douglas Futuyma (2010)

In 2005, a group of scientists from La Trobe University in Australia investigated how species will adapt to global warming by studying a species of rainforest fly, Drosophila birchii (later published in Science).  Increased temperatures may lead to drier conditions in rainforests, so the authors wanted to see how quickly this fly could adapt and develop resistance to desiccation.

As in many directed evolution experiments, they took the flies to a lab and exposed them to dry conditions, with most of the flies unable to survive the dryness.  The survivors were then bred to one another.  Thus generation after generation, only the most desiccation resistant flies would survive, resulting in a population that could survive dry conditions potentially induced by global warming.  Right?

Well, that was the idea at least.  At first, their population of flies did show a bit of increased resistance.  However, in the proceeding 28 generations of selection, there was no increased resistance to desiccation, despite that previous papers had found increased resistance in 3 other Drosophila species.  Drosophila birchii was unable to adapt to these conditions.  What went wrong?

The assumptions of natural selection

Natural selection itself is based on three assumptions in a population.  The first is that there will be variation in traits, such as multiple colors of eyes or hair.  The second is that these traits be heritable through the generations, that children will inherit the traits of their parents.  The third is that these variable traits have differential fitness, or that some versions of a trait might help you survive better than another.  Thus certain trait variants will help its carrier organism survive better, passing that trait to its offspring which will in turn bear this trait.

In order for populations to adapt by natural selection, these three requirements must be fulfilled.  When a biologist sees any population, he or she typically assumes that these they are met, and I can’t really blame them.  All cellular life we know of on this planet has a hereditary mechanism, the gene, which has differential fitness depending on the variation, thus meeting requirements two and three.

But what about requirement one, genetic variation in a population?  It has frequently been assumed to also exist in all populations.  In his 2010 review, Douglas Futuyma quotes the famous geneticist Richard Lewontin, who concluded in his 1974 book that “genetic variation relevant to all aspects of the organism’s development and physiology exists in natural populations,” for “[t]here appears to be no character—morphogenetic, behavioral, physiological or cytological—that cannot be selected in Drosophila.”

The idea here is that at any point in a healthy population, there will be many variants of genes in the population, which are replenished infrequently by new mutations.  Thus, when selective pressures come around, there is a full stock of variants to differentially survive and lead to new adaptations.  This is why small populations and inbreeding are considered such problems: the small gene pool means fewer gene variants and an increased inability to adapt to new conditions.

But why should this have to be true?  Is a lab-bred stock of Drosophila really enough to show that there is variation in natural populations, when the lab has a completely different set of selective pressures?

Selection bias in evolutionary studies

When the researchers studying the rainforest Drosophila birchii analyzed the genetic diversity of their collected populations, they found very low variation in the desiccation gene group compared to other genes in the population such as wing size.  This lack of variation prevented the flies from adapting to new conditions – though the researchers weren’t sure why.  Perhaps these dessication genes had (relatively) recently been under selective pressure, and had not had time to reconstitute the genetic diversity in the population.

Unfortunately, there are very few studies of this kind so it’s hard to draw any conclusions.  After all, people studying evolution usually want to study EVOLUTION IN ACTION and not evolution when it doesn’t happen.  I suspect that past experiments that have found low genetic diversity have been tossed simply because it seems less interesting evolutionarily – when maybe it is in fact more interesting.

This would be a form of selection bias: the choice of study populations or species unconsciously (or consciously) swayed by the desired outcome of EVOLUTION IN ACTION.  In his 1991 Croonian Lecture at the at the Royal Society, Anthony Bradshaw gave a compelling example of this sort of selection bias in evolutionary studies (published here).

In Prescot, Merseyside, a copper refinery opened up next to a meadow that had many species of grasses and wildflowers.  Over time, the soil was contaminated by copper, killing off many of the plants.  However, after 70 years, there were 5 species that were still able to thrive in these meadows!  They had adapted and developed a resistance to copper.

Even now I’m thinking to myself, “oh! cool!  How did they adapt that way?  How did that mechanism work?  How quickly did the resistance gene spread?” etc.  But what I’m forgetting is that, while five species did adapt, twenty-one species failed to adapt.  If there really was massive diversity at all genes in each population, you would think that at least one would confer some benefit to survive the copper.  But this did not happen: twenty-one species went locally extinct.

Just as interesting a question as “How did these 5 adapt?” is “Why did these 21 fail to adapt?”  But it’s a question that’s only begun to be reconsidered.  Death and extinction are far more powerful forces in shaping the whole of biodiversity on our planet than successful adaptation, but these evolutionary failures that occur all around us are little studied.  Especially as we anticipate global climate change causing unknown impacts on species worldwide, we should be studying non-evolution to get a better sense of what natural genetic variation actually looks like and what we may be facing in the future.

Literature Cited

Bradshaw, A. (1991). The Croonian Lecture, 1991: Genostasis and the Limits to Evolution Philosophical Transactions of the Royal Society B: Biological Sciences, 333 (1267), 289-305 DOI: 10.1098/rstb.1991.0079

Futuyma, D. (2010). EVOLUTIONARY CONSTRAINT AND ECOLOGICAL CONSEQUENCES Evolution, 64 (7), 1865-1884 DOI: 10.1111/j.1558-5646.2010.00960.x

Hoffmann, A. (2003). Low Potential for Climatic Stress Adaptation in a Rainforest Drosophila Species Science, 301 (5629), 100-102 DOI: 10.1126/science.1084296

Written by Hanner

January 5, 2011 at 10:58 am

28 Responses

Subscribe to comments with RSS.

  1. […] When adaptation doesn’t happen […]

  2. This concept is kind of similar to the punctuated equilibrium by S J Gould form the standpoint that you are looking at what is NOT happening (adaptation in one, slow progressive mutation in the other). We human beings do have a tendency to ignore the negative/lack of anything. Maybe it has to do with our pattern recognition tendencies?
    Very interesting.

    Yannis Guerra

    January 5, 2011 at 12:38 pm

  3. Good post, and raises an important point for consideration in any ‘direct evolution’ studies. When working with RNA or peptide aptamers, researchers will be aware of the potential variation (e.g. 1015 variations), but also aware that in the experimental design you can only really hope to see a modest representation of these (104-106), even with multiple iterations.

    There is a certain carry over to a biological entities, I try to get bacteria to become resistant to antibiotics to test both the frequency of incidence of this, and what effect it has on the evolutionary fitness of the bug; however, one thing we find is that its the environment (in vitro, in vivo, in situ) under which these experiments are performed that have the greatest effect on the outcome.

    Incidentally, I would have been interested to see whether information about evolutionary canalisation, and the nice (but yet to be unequivocally demonstrated) hypothesis of evolutionary capacitance; the idea of ‘locked in’ or suppressed variation that can be ‘released’ under suitable environmental perturbation.


    January 5, 2011 at 12:42 pm

  4. No offense, but this is crap. It’s well-known that the amount of genetic variation in populations is a function of population size and inbreeding. No one has suggested that all populations have the same amount of variation since (at least) the 1930s, and definitely not since people began sequencing genes.

    Yes, studying the level of genetic variation in natural populations is hard, because we don’t have an infinite supply of money and graduate students. But the fact that those guys “aren’t sure why” variation is lower in dessication genes in Drosophila isn’t a function of lack of understanding on this issue. If, in fact, there was a selective sweep, the signature should be pretty damn easy to identify to people who know what they’re doing, if they can muster the resources to do the necessary follow-up. If they’re unable to, it’s probably only because they don’t have millions of dollars to spend sequencing SNPs in their wild population.

    I’m surprised that you seem to be unaware of the copious literature connecting diversity to selection and population size. Pretty much every basic population genetics text book goes over this ground.


    January 5, 2011 at 12:45 pm

  5. …hmm, and that should read 10E15, and 10E4-10E6 (superscript tag not working).


    January 5, 2011 at 12:45 pm

  6. You made BoingBoing! This is certainly a benchmark of some kind, and should be celebrated, probably.

    I started thinking about this last week, actually, as I was listening to Radiolab. They were playing a clip about the natural state of starvation and suffering and how essential that is to the evolutionary process. This led to a discussion of altruism as a force of evolutionary pressure, which I guess is somewhat related in that it isn’t a traditional selective pressure.

    Maybe we are looking at a three-headed beast of evolutionary pressure: The success of adaptation in response to selective pressure, the rate of failure of that adaptation, and then you add in altruism and sacrifice for the good of the population . . . this is a fun topic.


    January 5, 2011 at 1:05 pm

  7. Just as interesting a question as “How did these 5 adapt?” is “Why did these 21 fail to adapt?” But it’s a question that’s only begun to be reconsidered. Death and extinction are far more powerful forces in shaping the whole of biodiversity on our planet than successful adaptation, but these evolutionary failures that occur all around us are little studied.

    I agree. I think one needs to ask both questions to get the fullest understanding. The same goes for every extinction event. A just as interesting (if not more interesting) question than “What killed off the dinosaurs?” is “What made birds, placentals, ants, etc. survive the K-T event?” But the bigger picture requires that you get answers to both questions.

    The Science Pundit

    January 5, 2011 at 1:11 pm

  8. Hey Hannah – first, my name is Saurabh, just to correct.

    I apologize if I misread something; in your discussion of population size you say: “Is a lab-bred stock of Drosophila really enough to show that there is variation in natural populations, when the lab has a completely different set of selective pressures?” Perhaps you’re unaware of how genetic variation is maintained, or the high level of variation in (lab-bred) Drosophila. It is pretty much impossible that lab-bred Drosophila could have more genetic than the wild population.

    Futuyama gives an account of theories for why adaptation should fail. They include low levels of genetic variation, yes (in certain locally-adapted “specialist” subpopulations), but also include things like smaller mutation target for trait resistance, trait combinations, weird sexual selection effects and stuff that is far more complicated and interesting than just a simple lack of diversity in the population.


    January 5, 2011 at 1:41 pm

    • Hey Saurabh,
      Sorry about the name. I’m glad you have your conversation tone on now! Yay!

      You are right – there are many reasons that are a bit more complicated than simple “genetic variation” but I felt that going into that kind of detail would ensure that the point I was trying to make would get lost. Things like pleiotropy and epistasis in particular are huge hurdles that cannot be explained by lack of genetic variation alone, as well as ecological factors. Maybe I oversimplified a bit. (Although I do still stand that populations can have low variation, even if it is just a remnant from selection in previous millions of years.)

      In the end, I just wanted to make clear that, no matter what the mechanism of failure, studying these evolutionary failures is interesting and worth our time.

      Hannah Waters

      January 5, 2011 at 2:13 pm

  9. I’m just curious what effect sexual reproduction has on these results. I’m not an expert on flies, but humans tend to reproduce in the age range of 15 – 35 years. Those humans who survive the longest (into the 100’s) are not reproducing at that age. The population sample that is reproducing at an earlier age can’t be determined if they will be the ones most likely to survive the harsh conditions, since at 15 – 35 everyone is quite hardy (or horny).

    So, with the flies, were only the ones that survived the longest only then allowed to sexually reproduce and create subsequent generations? That would seem to me the only way to ensure only the genes of the long term survivors made it so subsequent generations.


    January 5, 2011 at 2:55 pm

  10. I think you want to relocate your copper facility to Prescot, Merseyside, in England from Maryland.

    More importantly, it seems from the lecture that not only did those 21 species become locally extinct, but that the five species that remained were not present locally outside the immediate area of the facility. Could these have been already-copper-evolved species which were uncompetitive in the area for other reasons and their airborne seeds were able to take root only because of the ebb of the other species?


    January 5, 2011 at 4:16 pm

    • Hey Greg,

      You’re certainly right – it’s not only genetics that determines a species’ competitive ability, but also ecology and community structure. I don’t have a full answer to your question – I’m not sure how common copper-resistance is generally, or whether the plants in question would have undergone pressure towards copper resistance previously.

      Great question though. Just one reason why it is hard to study evolution in the field – many other factors at work.

      Thanks for reading,

      PS: Thanks for the location correction!

      Hannah Waters

      January 9, 2011 at 12:14 am

  11. […] When adaptation doesn’t happen “Evolutionary biology has been enriched by considering not only how adaptation happens, but also why it often does […] […]

  12. […] are far more powerful forces in shaping the whole of biodiversity on our planet than successful adaptation, but these evolutionary failures that occur all around us are little studied.  Especially as we […]

    Grid Types

    January 5, 2011 at 8:23 pm

  13. I like the use of Game Theory in Biology. It is very concrete and logical, especially when applied to it through Evolutionarily Stable Strategy (a Nash Equillibrium refinement.


    January 5, 2011 at 8:48 pm

  14. Right, researchers studying evolution tend to find evolution and ignore non-evolution. The fact we know so little about the 21 failed adapters shouldn’t be surprising. Most researchers freak out if their work doesn’t validate the status quo and will “interpret” their data to fit the accepted belief system. It takes a real brave person to tackle the non-conformists.

    BTW, cool blog!


    January 6, 2011 at 1:12 am

  15. Great post – I’ll be sharing this with my high school students. And btw, I was thinking along the same lines as Greg, re: the 5 surviving species near the copper plant.


    January 6, 2011 at 8:56 am

  16. […] When adaptation doesn’t happen | Culturing Science – biology as relevant to us earthly b… Just as interesting a question as “How did these 5 adapt?” is “Why did these 21 fail to adapt?” (tagged: evolution science biology research ) […]

  17. Nice work,

    I think the other interesting assumption that is lurking in the “adaptation will happen” assumption is that there are multiple loci underlying most traits. So, in The Growth of Evolutionary Thought Mayr takes the Lewontin etc experiments of teh 70s to show argue that variation is almost limitless, but then argues it mainly arises from recombination of existing genes rather the mutation. It appears in the flies there was only one major locus with any genetic variance for desiccation resistance, once that was swept to fixation there was no variance for selection to act on.

    Will have to have a read of the Futuyma paper to see about these more esoteric ways that adaptation fails to happen.

    david winter

    January 9, 2011 at 12:01 am

    • Hey David,

      Futuyma spends time on pleiotropy and epistasis, locus size and location, among other things that I cannot list off the top of my head. :)

      I do wish I had taken more evolution classes before trying to write such broad posts like this one – so many “unknown unknowns” that I feel I have overlooked. I have a lot of reading to do…

      I have a few other reviews that I didn’t directly cite in here if you want more!


      Hannah Waters

      January 9, 2011 at 12:18 am

  18. […] You can read the full post HERE. […]

  19. Interesting post Hannah.

    I wonder whether the failure to adapt is necessarily due to a lack of variation. When all organisms in an ecosystem are being challenged by significant changes some of them may be competing for the same niche in the new environment. So the extinction events may just come down to niche competition, rather than absence of sufficient variation.

    John Normal

    January 10, 2011 at 12:40 pm

  20. Global warming causes increased rainfall, not desertification, *sigh*. Earth’s deserts shrink in times of warmer climate.

    Simon Newman

    January 11, 2011 at 5:55 am

  21. Reading about the experiment I do find myself wondering of what value it has in evolutionary science.

    Tracking the genes of rapidly reproducing organisms may provide the ability to perform classical experiments, but it ignores that evolution is far wider. Unless predation is introduced, or the ability of the given species to find new enviroments or food sources then the results may well have usefulness to biology as a whole, but be of little understanding of evolution.

    Indeed it may well be as a dangerous to study of the subject as the philosophical misuse of Darwin by German Militarists, eugenisists and Soviet collectivists in the early 20th century.


    January 12, 2011 at 9:27 am

  22. […] Culturing Science: When adaptation doesn’t happen, and from Laelaps: The Pelican’s […]

  23. […] of the day’s news at Discover, Hannah Waters who won an award of Best New Blog last year and it’s not hard to see why, and new blogs by seasoned journos like Claire Ainsworth, writing on […]

  24. Hi,
    Why does the question arise “When adaptation doesn’t happen?” Actually adaptation seldom happens. With so many fossils of animals and plants which have gone extinct it tell us that adaptation is not a process which occurs every time and everywhere. For adaptation to occur there has to be few more things apart from variations present. And the most important of the factors is time. If a species get time enough then the variations can arise but if the selection pressure is too great but variation small then the species is bound to extinct. Given time any species would. Adaptation time occurring ever time when a selction pressure is applied is a myth.


    April 19, 2012 at 3:10 am

Comments are closed.

%d bloggers like this: