Funny.
If we assume that the distribution is measuring some trait (e.g. "testosterone content," "femininity," measured however you will), and it's bimodal (distribution is dominated by two binary sexes), then there will be people on either side of both peaks.