The incidence of divorce within cohorts of American marriages contracted since the Civil War.

The proportion of marriages that end in divorce can be estimated from vital registration or from census data. The former source suggests considerably higher levels of divorce than does the latter. A new series, combinig the two sources, is presented for annual marriage cohorts back to 1867. Actual experience to 1970 is traced and a projection beyond that point is made for cohorts with incomplete divorce histories. The cohort divorce series moves steadily upwards and shows much less variability than an equivalent series of period divorce rates. Factors related to high divorce within a cohort are armed service mobilization and high unemployment rates in the year of marriage, and slow national economic growth between pre- and post-marital periods.