There is general agreement that adaptive variation of sex ratio at birth has not been decisively demonstrated in primates (including human beings). So some workers have questioned whether it actually exists. Others have conjectured that it exists but is subject to as yet unidentified 'constraints' (factors opposing the modifying influences of selection in the phenotype). Meanwhile though most workers have called for research to reveal the proximate causes of sex ratio variation, few (if any) have directed studies toward that end. Here it is argued that hormonal action is responsible both for much adaptive and non-adaptive sex ratio variation, and for constraints on the adaptive variation. My hypothesis proposes that levels of steroid hormones (testosterone and oestrogen) of both parents around the time of conception are positively associated with offspring sex ratio (proportion male at birth) of mammals including man. Testosterone in men and oestrogen in women are also known to be positively associated with the health, attractiveness and fertility of individual human beings. However, high levels of testosterone in women are frequently associated with adverse medical conditions. It is suggested that for these reasons (and contrary to some adaptive theory) some classes of people (particularly women) in suboptimal health ("condition") produce excesses of sons. It seems that gonadal hormones are responsible for adaptive variation; and that maternal adrenal hormones are responsible for maladaptive variation. In evolutionary terms, gonadal hormones precede adrenal hormones.