>>12082702OK, a small thought experiment. Since PC1 accounts for ~80% variance and PC2 for ~10%, we're left wondering what the hell PC1 actually encodes. What if PC1 corresponds to skin color (the most obvious difference between the listed ethnicities), and PC2 is some rather minor facial features where Africans appear as right in the middle of the very smoothly distributed ethnicities, and the rest of the PCs are some minor visible or invisible diffierences. Without the loadings in the PCA it's impossible to tell what it might actually code for, and as such the results are entirely useless, and any conclusion from them will be baseless extrapolation.
