Key concepts

Figure 1. Ecological geneticists often use common garden experiments to understand the potential determinants of phenotypic variation in natural populations. Here, I assume two populations of tadpoles. Population A lives in a habitat devoid of predators and Population B lives in a habitat with predaceous dragonflies, creating a potential landscape mosaic of heterogeneous selection. The two tadpoles differ in their phenotypes such that the ones living with the predaceous dragonflies have larger tailfins than the other population. A common garden experiment is performed in which eggs from each population are collected from the two natural populations and raised in a controlled environment with and without dragonflies (with each combination replicated many times). Three divergent phenotypic outcomes that might characterize tadpoles raised in the common garden are depicted. If individuals from both populations have bigger tailfins when grown with predaceous dragonflies, but otherwise the populations have similar phenotypes, then nongenetic phenotypic plasticity likely underlies the observed variation. However, if Population B consistently has larger tailfins regardless of treatment and Population A does not (2nd row), then this finding suggests that bigger tailfins might have evolved in Population B. Two caveats are worth mentioning here. First, we would need to confirm that bigger tailfins are associated with higher fitness under attack by predaceous dragonflies through a natural selection experiment. If tailfins do not give rise to higher fitness, they might have evolved due to random drift. Second, maternal effects, the nongenetic inheritance of traits from mothers (e.g. bigger mothers have bigger offspring) could confound results; whenever possible, several generations should be raised in the common garden to eliminate maternal effects. In the topmost row, Population B shows a plastic reaction to dragonflies, whereas Population A does not. This result suggests the evolution of plasticity (a gene by environment interaction) in Population A.

Figure 2. Adaptive evolution can occur quite quickly, demonstrating substantial changes in trait evolution over the course of just a few generations. Here, the median rates of adaptation (in standard deviations of phenotypic change per generation, the ‘Haldane’) compiled from published studies are arranged by geometrically distributed bins. Each bin estimates the evolutionary rates from published studies analysed over the number of generations greater than the prior bin up to and including the value of the bin. Note that no data were available for the 1–2 generations bin. Evolutionary rates are high at first but decline when analysed over longer time spans, possibly because optima are quickly reached or because natural selection gradients rapidly reverse, leading to less overall long-term phenotypic change. Data from Kinnison and Hendry (2001).