Using 1995-2015 Dutch marriage registry data, we show that daughters are associated with higher divorce risks, but only when they are 13 to 18 years old. There are no detectable gender differences before or after those ages. These age-specific findings are at odds with son-preference and selection explanations for differences in divorce risks. Instead, the findings point to explanations which involve family relationship dynamics associated with teenage sons and daughters. We find supporting evidence of relationship explanations in supplemental analyses of Dutch survey data. We also find that teenage daughters are associated with higher divorce in the U.S. in analyses of the Current Population Survey Marriage and Fertility Supplements.