A letter and a response were published online in the American Journal of Physical Anthropology (no abstracts). Skin-color sexual dimorphism is the lighter pigmentation of women (of a population) relative to the men.
Human skin-color sexual dimorphism: A test of the sexual selection hypothesis. Reply to Frost (2007) (p NA)
Lorena Madrigal, William Kelly
Human skin-color sexual dimorphism: A test of the sexual selection hypothesis (p NA)
Women have lighter skin than men do across a wide range of populations, even on the unexposed skin of the upper inner arm, possibly because of sexual selection by men for lighter-skinned women. If this hypothesis is true, human skin color should become more sexually dimorphic with increasing distance from the equator, since sexual selection for lighter skin in women would be less constrained by natural selection for darker skin in both sexes. Yet when Madrigal and Kelly (2006) analyzed skin reflectance data from 53 different samples, they found that the most dimorphic human populations were actually those of medium skin color at medium latitudes.
There is a discussion about this issue in the older blog post about Madrigal and Kelly's study with some comments from Peter Frost himself. That article also contains a very useful table of skin reflectance data in the end. Some Caucasoid (incl. some Indian ones) data points from the table (at wavelength 685), with females first, males second, and higher values designating lighter skin (more reflectance). In these data points it looks to me that Iranians and Kurds have the highest sexual dimorphism. Are Iranian males really much darker than females?
Punjab 53.64 54.52
Sikh 55.52 53.2
Jirel 56.64 49.12
Yemen 56.94 53.15
Algiers 58.04 58
Iran 60.1 51.2
Kurdish 60.02 54.89
Europeans 63.1 61.5
Belgians 63.65 64.51
Viscaya (Basques) 65.78 65.25
Belgians 65.9 67.3
Guipuzcoa (Basques) 66.38 65.53