A quick mea culpa. Yesterday I put up a post on the difference in height between northern and southern Europe, following the lead of the heading of the paper which I blogged. But, in the text they do note that their sample is skewed toward northern Europe. Additionally, their geographic coverage is stated in the supplements. As noted by some commenters not only is it northern Europe skewed, but it’s really western Europe biased. There’s nothing wrong with that as such, but it leaves much of Europe outside of this west-central transect unsampled. Therefore, I’m a little more cautious of making pan-European latitudinal generalizations.
That being said, I still suspect there is going to be spatially structured differences in the concentration of alleles which predispose one to great height. I’d especially be curious to see if the people of the Dinaric region tend to cluster with northern Europeans, rather than their Balkan neighbors. Please note that one of the important aspects of this study is that they replicated their findings among siblings. When observing correlations between traits on a population-by-population basis and then extrapolating, it is of the essence that those patterns can be replicated in family-based studies. This applies to within population observations as well. For example, there is some correlation between height and intelligence. But that correlation disappears among siblings (i.e., tall siblings are no more intelligent than short siblings).
Finally, a friend brought to my attention some serious concerns about the evolutionary quantitative genetic model outlined in the paper. After reading their critique I would say that though they have convinced me of the likely importance of these alleles in generating inter-population differences, I am less than confident of their adaptive model.