Havelock Ellis on British brunettes

I had previously written about Havelock Ellis' study at the National Portrait Gallery. Recently, I came across an interesting excerpt by Ellis, which confirms my impression that brunettes are overrepresented in the British aristocracy, and gives an evolutionary explanation for this phenomenon. His explanation ties in quite well with my theory on the evolution of blondness. In summary, my theory holds that:

1. Hair acts like a frame, emphasizing facial features.
   
2. Light hair "softens" features, while dark hair makes them more prominent, just as a frame draws attention to the painting.
   
3. Women with "hard" features have a higher reproductive fitness if they are blond, since they look more feminine, due to the above-mentioned softening effect.
   
4. Hence, there is sexual selection in favor of light hair which however operates primarily on facially robust and un-feminine populations.
   
5. This selection allows the robust and un-feminine characteristics to persist if they are linked with light hair; by contrast, in a brunet population, sexual selection operates on facial characteristics alone, leading to a swifter selection for facial beauty.
   
6. Superior beauty is primarily a function of facial features, and since (i) good facial features are enhanced by framing, and (ii) brunet populations evolve towards a more attractive facial structure faster, it follows that brunettes are on average more attractive than blondes facially.
   

Of course, I personally accept the last point of the above list, but it is possible that I might be biased. I had previously discovered a BBC Story which seemed to confirm my opinion. It was thus pleasantly surprising to see that this preference for dark hair is nothing new, but was noticed almost a century ago, and indeed by Havelock Ellis, whose escapade in the National Portrait Gallery was intended to show the alleged "superiority" of blonds.

I came across this quote in Joana Pitman's rather fluffy book On Blondes, p. 168, citing Ellis in The Monthly Review, August 1901. Ellis' words are in quotes (')

The low position of the hereditary aristocracy [DP: low in "blondness"] was also explained by infusions of foreign blood, and because 'peers have been in a position to select as wives... the most beautiful women, and there can be little doubt that the most beautiful women, at all events in our own country, have tended more to be dark than to be fair. This is proved by the low index of pigmentation of the famous beauties in the Gallery.'

It would certainly be interesting to do an experiment that would confirm this theory. I can think of two possible setups:

1. Artificially change the pigmentation of women photos, and determine whether attractiveness ratings improve when unattractive brunettes are given light hair and deteriorate when unattractive blondes are given dark hair, and that no or the opposite effect occurs from attractive women of either color.
   
2. Compare random samples of blondes from a naturally blond population (e.g., from Sweden) with a random sample of brunettes from a naturally brunet one (e.g., from Spain) after removing the differences in hair color.