Scientists again see natural selection in real time

June 8, 2009 • 3:32 pm

In WEIT, I describe several studies showing that natural selection can change species over periods shorter than a human lifetime.  These studies have been important in convincing skeptics (although not creationists, who will never be convinced) that natural selection is more than just a speculation, but a force that can really mold animal behaviors, appearances, and reproductive traits in real time.  The most famous of these, of course, is the work of Peter and Rosemary Grant on a species of Darwin’s finch.  Their work showed that individuals’ beak and body characteristics could change over a single generation when an El Niño event changed the finch’s food.  (See The Beak of the Finch, by Jon Weiner, for a popular account of the Grants’ work.)

Now a group of researchers from Canada and the US have reported natural selection acting in guppies, molding their reproductive behavior over a period of only 8 years: roughly 13-26 guppy generations.  This work is based on the observation that in Trinidad, the common guppy (Poecilia reticulata) lives in streams both above and below waterfalls.  The guppy’s downstream predators cannot migrate past the waterfall barriers, so upstream guppies experience different environments from downstream guppies. In other words, downstream populations have to deal with much stronger predation.

This leads to some evolutionary predictions.  The body of evolutionary theory called “life history theory” predicts that when an animal species experiences higher predation, it should evolve a different reproductive strategy.  In short, guppies harassed by predators should reproduce earlier than un-predated guppies, for the greater your chance of being chomped, the less likely you are to leave offspring if you delay reproduction. Second, fish that experience less predation should produce larger embryos, since they have the luxury of delaying reproduction and because larger embryos make the newly hatched fish more competitive.  Finally, low-predation fish should have fewer offspring in each bout of reproduction, for it is to their advantage to spread their given lifetime allotment of reproductive effort over a longer time.

In 1996, some of these authors introduced guppies from downstream, high-predation populations into uninhabited upstream, above-waterfall areas. Eight years later, in 2004, they took samples from both ancestral and derived populations to see if any differences in life history had evolved.  And they did — in the predicted direction.  Upstream guppies had fewer but larger embryos than downstream guppies, as well as a small reproductive allotment (egg mass as a proportion of body mass).  Mark-recapture experiments on adults also showed that, when tested in the upstream environment, upstream-evolved guppies had higher survival than their downstream ancestors.

The analysis is arcane, but the results are clear:  guppies have changed their life histories in an adaptive way in only eight years.  This certainly reflects the action of natural selection, since previous studies have found similar results for life history, and also for color. (Guppies introduced to a low-predation regime evolve brighter colors in males; brightly colored males are favored everywhere by sexual selection but become disadvantageous in high-predator environments since they are more likely to be spotted and eaten.)

In toto, the guppy work is as powerful a body of evidence for selection, if not more so, than the work on finches.  This is not to denigrate the finch study, which is brilliant.  That work, however, was an uncontrolled “natural experiment” that affected two characters (bill and body size), while the guppy work has involved many groups of investigators doing controlled introductions — and all finding the predicted evolutionary changes in many characters.  Birds, of course, are more charismatic, but the humble guppy has a lot to show us about evolution.


S. P. Gordon et al. 2009.  Adaptive changes in life history and survival following a new guppy introduction.  The American Naturalist, Volume 174, pp. 34-45.


5 thoughts on “Scientists again see natural selection in real time

  1. Just wondering: did the two populations diverge enough to let speciation take place? If not, are there any ways to predict how many generations that might take?

  2. I’m behind the firewall at home this evening and so don’t have ready access to the Am Nat pdf. I want to look at it to see if the authors address one of the predictions whose accuracy is not summarized in the WEIT post (and it also goes to Deen’s question, above): Did reproductive timing change as predicted by the life history hypothesis (i.e., the downstream guppies should evolve to delay reproduction over time from transfer to upstream pools)?

    If so, such a divergence in reproductive strategy could readily lead to reproductive isolation and hence incipient speciation.

    Another point– and I apologize for the unsolicited (and likely unwarranted) “peer”-review: Did the authors perform reciprocal transplantation (upstream to downstream and the reverse) of guppy populations? It would seem that equivalent but opposite changes in life history upon reciprocal transfer would make quite a compelling story.

  3. I seem to recall some very similar studies done on guppies in an artificial laboratory environment (wish I had the citations, but I seem to have misplaced them…). It’s good to see those results reproduced in a wild environment.

    Also, this type of research is exactly the kind of research that needs to be done to test adaptive hypotheses. When you’ve got evidence like this, you have more than a “just so” story. If we could (cautiously!) extrapolate ideas based on this kind of research to human populations, we might be able to develop a robust evolutionary psychology, rather than the mostly speculative guess-work that goes on in EP these days.

  4. Just to clarify, here is an observation that would refute the Creationist view pretty decisively, and would also, I think, lead many people to abandon it:

    In a natural or controlled experiment like the above, scientists observe one population of conspecifics become geographically split into two, gradually drift apart in characteristics, and eventually form two visually distinguishable populations that cannot interbreed.

    1. I don’t understand why you wrote ‘would’ in italics. If this experiment is the answer then the answer is already present and not a futuristic stance.

      Also, I’m on the same page as Deen. When would we see a species change? Why don’t they speed that up in a lab to get the real evidence that is needed?

      The fish can still breed with each other. They just made some changes to how often and how many eggs they lie. This article shows some environmental adaptation but does not show how species are connected. Almost all living things change is some manner but it is the huge jumps that show holes in evolution.

Leave a Reply