The theory of evolution works on the principle that creatures adapt to environments, both geographical and situational (i.e. predators), as the species goes through generation after generation, correct?

With that in mind, how many generations would any given species have to endure before any evolutionary alteration is noticeable enough to be relevant?

The amount of change per generation is actually a function of how strong the selection is, and how much genetic variation there is in the population. in basic terms the equation is

evolutionary change = strength of selection x genetic variation

When we apply strong selection to populations in experiements we commonly get detectable change over a single generation. Typically however, selection experiments in biology aim for 5-10 generations.

Last edited by Alistair Wilson (17th May 2016 18:43:02)

agreed but that's in a lab/controlled environment. I suspect Taylor's question was related to a real world setting. other than very substantial environmental changes wouldn't it take longer than 5 generations - say for darwin's finches beak changes or human lightening of skin as we migrated north? Empirically I would have guessed at 50-100 generations if not more (but then I'm not an evolutionary biologist).

Ah. So the same conditions apply - in that it depends on the strength of selection and amount of genetic variance. However, temporal trends in traits in nature are also determined by changes in the environment which cannot be controlled in the field. So the first challenge for scientists is to try and determine the relative contributions of genetic change (=evolution) and environmental change to what is happening.

For some simple Mendelian traits this is trivial because of an (almost) 1:1 mapping between a genotype we can determine and a trait. In such cases we can literally measures gene frequency changes from 1 generation to the next and see the corresponding trait change. To robustly infer that these changes are adaptive evolution we strictly need to a) understand selection on the trait to make sure changes are in the expected direction and b) rule out genetic "drift" in the absence of selection as an alternative explanation for the change. Note change through drift IS most certainly evolution, but it is not "adaptive evolution" (i.e. response to natural selection) which is I think what Taylor is referring to specifically. Our ability to meet this criteria increases with the number of generations so while we may be able to see change over a single (or few) generations, we may not be able to rule out drift on statistical grounds until we have followed the genotype/trait frequency for longer (e.g. tens of generations).

For "complex" or polygenic traits - those like body weight or parental care behaviour which are likely to depend on many genes of small effect AND many environmental influences, "proving" adaptive evolution is real time is harder within a single population. This is because we can't simply track changing gene frequencies, and thus can never 100% exclude the possibility that some environmental variable we have not measured (and so can't control for) is driving the trait change. We can - and do - use statistical predictions of something called "genetic merit" which is essentially the expected trait value of an individual given everything we know about its genes, and see average merit is changing across generations. Studies to date have mostly done this is 10-20 generations and found statistical evidence for genetic change underpinning trait changes that we can see (and detect easily).

Last edited by Alistair Wilson (17th May 2016 18:42:26)

many thanks Alistair - that is very clearly put!

Some things can happen almost instantaneously - plants often spontaneously double their number of chromosomes when reproducting leading to offspring with twice the normal DNA. This means that they can't normally breed back with the parent species and may have large cells and be especially large themselves. In short, a single generation may be enough in some circumstances.