Log-linear analysis was used to investigate the relationships between spouse isonymy and parental isonymy, time, endogamy/exogamy and groom's occupation in a total of 6266 first marriages of parish-born spouses in the civil parishes of Harris, Barra and Stornoway, between 1855 and 1990. Stornoway was divided into rural and urban (Burgh) components.
Spouse isonymy comprised between 6% and 12% of all marriages; parental isonymy was generally slightly more frequent, due probably to the occurrence of occasional families in which both sets of parents were isonymous. Between 38% and 51% of marriages were endogamous.
In the multi-way contingency tables, good fits between observed and expected cell numbers were obtained by a simple model, comprising the main effect of spouse isonymy and the first-order interactions between spouse isonymy and the independent variables. Inspection of the coefficients showed that in all four populations spouse isonymy was positively related to parental isonymy (doubtfully so in the case of Barra). The other independent variables showed less clear-cut relationships-endogamy/exogamy in Barra, and groom's occupation in Stornoway Rural; time period showed no significant relationships.
It is concluded that the occurrence of this 'two-generational marital isonymy' might have an effect on the breeding structure of the population.