Attractive women have high estrogen?

This BBC story covers this paper (warning! PDF!) that found a correlation (r = 0.48) between attractiveness and estrogen level in women:

The findings make evolutionary sense - men are attracted to the most fertile women, the University of St Andrews team told a Royal Society journal.
Oestrogen levels during puberty can impact on appearance by affecting bone growth and skin texture, they said.
...
The team of psychologists at the University's Perception Lab photographed 59 young women's faces aged between 18 and 25 and analysed their sex hormone levels.
They then asked 30 volunteers - 15 male and 15 female - to rate the faces according to attractiveness.
Both male and female volunteers rated the faces of the women with the highest hormone levels as the most attractive.

My first reaction was that it just seems so...unlikely.

My more studied reaction is that I'm not sure that the results are as interesting as they look at first glance. The paper is conservative in its interpretations -- much more so than the comments in the BBC article would suggest.

The general idea is that males want to be able to assess female fertility, and if they can "read" estrogen levels through facial characteristics (which they perceive as "attractiveness") then it should have been adaptive.

This argument would have to be based on two correlations: a correlation between "attractiveness" and estrogen level during facial development, and a correlation of estrogen level during facial development and fertility. This paper reports the first of the two.

The second is (as far as I know) unknown, although there may be some hints in the opposite direction. There is this paper for example, describing the effect of estrogen treatment to reduce final stature in tall girls (estrogen promotes early epiphyseal fusion):

Fertility problems were more prevalent in women previously given estrogens to lessen their adult height....Among women attempting to become pregnant for the first time, the likelihood of conceiving every month was much lower for treated than for untreated women ( Fig. 1). Women treated for tall stature had a significantly lower age-adjusted per-cycle rate of conceiving a first pregnancy. Fecundability was impaired both in women treated with DES and those given EE, and the timing of treatment (before or after menarche) and its duration did not influence the findings.

It seems to me that if pubertal estrogen levels influence fertility, the strongest effect should be on the age at menarche -- since that influences the reproductive lifespan.

Menarche would be a much more important reason for men to be able to assess estrogen levels in developing women. In small human groups, most men probably would have had a very good knowledge of all the women who would be potential mates. Attractiveness might have made some considerable difference to mating decisions, but these would also have been subject to many other constraints. On the other hand, it would be of great value to be able to accurately assess when new maturing young women would become available to mating. A young woman would likely be well-served to provoke some competition between males. Signaling maturity through physical changes would be a way to spur this competition -- which in many cases would have played out over many months or even years.

A major influence on age at menarche is fatness, and fatness is related to estrogen levels in older women (Kaplowitz et al. 2001). Fatness may also significantly affect assessments of attractiveness, although not in an obvious way. I would guess that a slightly fuller face would be generally regarded as more "feminine", and in the study "feminine" ratings were correlated (r=0.97) with attractiveness. Certainly the average physiognomy of the 10 high-estrogen women in their sample has a fuller face than the average of the 10 low-estrogen women:

Left: 10 highest-estrogen average face. Center: 10 lowest-estrogen average. Right: Actress Lindsay Lohan giving come-hither look.

On the other hand, facial fullness -- especially in the lower face -- may not be a good guide to fatness generally. It's all just so confusing!

The pictures do give a good hint to the "make-up effect" that the article discusses:

"The findings about make-up are also interesting. The implication is that women are employing a deceptive strategy. They can fool the male visual system with make-up."

Yeah. They appear to fool "love bugs", too.