(I’m just kidding about “colour”; I switched my spell-checker from US English to UK English, though, so I have to keep alert to my u’s and zeds.)
Randall Munroe of the top-rank webcomic xkcd amassed a gargantuan dump of imperfect-but-fascinating data about colour perception and naming. It doesn’t bring to my mind any grand conclusions, doesn’t seem to prove anything about arbitrariness of colour identification or gender difference, but it’s fun to read about anyway — and I doubt that if one had adopted a means of removing the various sorts of bias from his data, you’d get results that differed appreciably. It’s a lovely bit of online experimentation and (on Munroe’s part) presentation of results. Cheers all around!