The NS article I referred to, McGrath, was quite specific in reporting that the investigators found the it was the hip:waist ratio or 1.2:1 that was the only 'measure' that stood up. It was so long ago now that I can't recall some of the other specifics but they certainly ruled out any single criterion such as waist size. And boob size.

Amos' post after mine reminds me that other NS articles, at around the same time as the one I summarised, looked at other putatively evolutionary aspects of attractiveness and its consequences. The series (NS never 'drew attention' to any notion of a 'series'; that was my doing) started with a report on DNA analyses of siblings and found there was a disproportionately (to them) high rate of families where the mother's husband was not the father of all her offspring even though, putatively, he ought to be. In exploring this phenomenon (from an evolutionary perspective, you understand) someone carried out an investigation in which women diarised their randiness, sexual activity, menstrual cyles as well as whether or not they were using contraception (and type) or not.

From memory, they found that women were randier and more sexually active at times when they were more likely to conceive and those that were in committed relationships who 'played away' were more likely to do so also at times when they were more likely to conceive. While diaries are not as objective as soom would wish, Amos' post above would seem to support hypotheses that men would find women more attractive at times when they're more likely to conceive, a corrollary to the first of the findings.

The second leads to an hypothesis that women may choose their social partner on the basis of his ability to support herself and dependants but also choose to widen the biological competition by requiring his sperm to be competitive as well.

The NS series went further but I'll leave it for now.

Cheers, Rowan