Fluctuating selection, canalisation, and evolvability: Why macroevolution is not “microevolution writ large”

One of my biggest pet peeves is when people say that macroevolution is just “microevolution writ large”; this is a common saying especially among the creationist-debunkers to counter the claim that microevolution happens but macroevolution doesn’t. It infuriates me to no end, particularly because one of my biggest research goals is to identify the factors affecting the gulf between micro- and macroevolution, or why short-term microevolutionary changes do not accumulate to fixated and exhibited macroevolutionary changes.

And I’m not the only one who does this research, as this remains one of the most open questions in evolutionary biology (Arnold et al., 2001). Part of the problem is that environments change all the time, and with them change the selective optima – what may be the most fit phenotype at one point in time might be the most unfit several generations later. The leads to a pattern of fluctuating selection, where the optimal phenotype(s) change every so often.

Fluctuating selection is generally agreed on as a factor in the evolution of evolvability (Wagner & Altenberg, 1996), the ability to maintain a high rate of microevolution, which in turn is dependent on the concept of canalisation first espoused by Waddington (1942). While he was talking about it in terms of development and evo-devo, we co-opt it in this discussion to talk about population genetics, since that’s the framework for microevolution. Wagner et al. (1997) define it under this framework as an effect leading to the lessening of the effect of mutations.

In other words, by decreasing canalisation, you increase evolvability, since evolvability demands the ability to produce new effectual mutations (Moreno, 1994). This then links back to fluctuating selection because in order to respond to a regime where the optimal phenotype changes constantly, higher evolvability and less canalisation is needed. Studies show exactly this: under fluctuating selection, decanalisation alleles and polymorphic genes become selected for (Kawecki, 2000), i.e. evolvability rises as expected.

An analogous way to consider this is by thinking about how parasites and other highly-specialised organisms (highly-canalised organisms) tend to reach evolutionary dead-ends, while generalists can always specialise further (highly evolvable).

That analogy brings us back to the phenotype and the bridge to macroevolution. Consider that microevolutionary changes, over many generations, are basically subject to cycles of fluctuating selection. If we generalise that fluctuating selection leads to higher evolvability and thus a higher amount of effective mutations, that means that the genotypes generated by generations of microevolution are geared towards producing variability, which then translates to a higher diversity of phenotypes. This is why we can’t say that macroevolution is microevolution writ large: by doing so, we imply that macroevolution is canalised by microevolution, when microevolution does not undergo canalisation except under strict and consistent selectionist regimes.

And even then, the implication that the genotype is a precise blueprint for a single morphology is false, since it ignores the various filters put in place by development and phenotypic plasticity. So, in all, please don’t say that microevolution is equal to macroevolution. They’re different fields (and they’re taught separately: any good evolution program will have a specific course on macroevolution and another on microevolution), they involve different phenomena, and the process of evolution is different at both levels. This is why we talk of a bridge between micro- and macroevolution: there is a chasm to be crossed. Macroevolution is not merely the magnified version of microevolution.


Arnold SJ, Pfender ME & Jones AG. 2001. The adaptive landscape as a conceptual bridge between micro- and macroevolution. Genetica 112-112, 9-32.

Kawecki TJ. 2000. The evolution of genetic canalization under fluctuating selection. Evolution 54, 1-12.

Moreno G. 1994. Genetic Architecture, Genetic Behavior, and Character Evolution. Annual Review of Ecology and Systematics 25, 31-44.

Waddington CH. 1942. Canalization of development and the inheritance of acquired characters. Nature 150, 563-565.

Wagner GP & Altenberg L. 1996. Perspective: Complex Adaptations and the Evolution of Evolvability. Evolution 50, 967-976.

Wagner GP, Booth G & Bagheri-Chaichian H. 1997. A Population Genetic Theory of Canalization. Evolution 51, 329-347.


  1. Hi Marc,

    Thanks for your post – very interesting stuff!

    I am curious though because you refer to Kawecki 2000 to support your statement that fluctuating selection favours de-canalizing genes. My interpretation of Kawecki 2000, however, is that the opposite is true. That is that weak to moderately strong fluctuating selection favours developmental modifiers that reduce the phenotypic expression of genetic variation (i.e. canalization).

    Best regards,

  2. Marc Srour

    Hey Julian,

    The reference to Kawecki 2000 should have been worded better, you’re right :) To me, the main result was that for long-term fluctuations (periodicities of 10+ generations), the result is polymorphism and decanalisation. The opposite effect happens only with quickly-fluctuating selection.

    I consider long-term fluctuations to be more frequent in practice, although I honestly can’t cite anything for that.


Leave a Reply