Assumptions about the 'typical' age gap between spouses underlie much social policy (e.g. the five-year difference in men's and women's state pension ages). In order to test the basis for these assumptions, detailed marriage registration statistics were obtained for 1963 and 1998, for England and Wales. Age differences between spouses were calculated and analysed by year, age at marriage and previous marital status. The median age gap hardly changed between 1963 and 1998 but this concealed considerable increase in the proportion of marriages where the man was younger than the woman or--to a lesser extent--where the man was six or more years older.