Document Type



This paper uses marital migration data transcribed from the Civil Registers of Marriage 1840–1911 to estimate kinship from migration matrices and isonymy in the Ards Peninsula, Northern Ireland. The distribution of religious denominations (Presbyterian, Episcopalian, and Roman Catholic) varies systematically throughout the region, with up to 77% Roman Catholic in the south and 81% Presbyterian in the north. Portavogie, a fishing village on the east coast, is exclusively Protestant, with a population 93% Presbyterian. Comparison of migration and isonymy with geographical distance by multidimensional scaling and the MATFIT procedure show Portavogie to be an outlier, more distantly related to other areas than its geographical position would predict. We suggest that this discrepancy is due to settlement history and occupational and religious isolation. Mantel tests show that marital migration is significantly related to geographical distance (rMG = 0.4257), as is the distribution of religious denominations (rRG = 0.5548) through settlement history. Migration is dependent on religion (rMR = 0.3674), and isonymy is dependent on migration (rIM = 0.2531) but not on geography or religion. With Portavogie omitted from the analysis, the dependence of migration on geography and on religion increases (rMG = 0.5583, rMR = 0.5646), as does the correlation between religion and geography (rRG = 0.7213). The dependence of isonymy on migration increases (rIM = 0.5103), and significant correlations between isonymy and religion (rIR = 0.4135) and isonymy and geography (rIG = 0.4660) appear. We argue that a full explanation of population structure requires geographical distance, settlement history, and the influence of religion and occupation to be taken into account.