Long term trends in mean phenotypes in wild animal populations have been of large interest, particularly in studies investigating evolutionary responses to harvesting or climate change driven changes in species and populations.
In their recent article Nine decades of decreasing phenotypic variability in Atlantic cod (Ecology Letters 2009, in press) Olsen et al. adopted a novel approach and not only analysed trends in mean size of juvenile cod but also investigated parallel changes in variability of body size. The outcome was very interesting: no shifts in mean size but a consistent, decreasing trend in variability of body length. A capture-mark-recapture study conducted over a subset of years revealed that the juveniles actually experience selection against the combinations of small body length & low growth rate, and large body length & high growth rate. This matched well with detected long term decrease in body length variability, suggesting that the selection pattern observed during recent years might have been prevailing for a longer time.
In the light of earlier studies focusing on mean phenotypes, the message clearly emerging from this study would be that the absence of a trend in a mean does not imply absence of evolution: it might just act hidden in the variance.
One Comment
Indeed. Also variance is something neglected in experimental studies for some reason.
Something different. If I recall right, the homogeneity of variances are the prerequisite of parametric tests, perhaps (considering large enough sample size) being far more important than normality. Can it be a reason why people do not force this thing too much?
One Trackback
[...] in this field have explored shifts in mean traits values. Earlier, I have blogged about a study by Olsen et al. (2009) challenging this view by demonstrating how fishing might have decreased phenotypic variance in cod [...]