This article uses isonymy to test predictions about the genetic structure of Irish populations made on the basis of geography and population history, and compares the mid-nineteenth century population of Ireland with the late nineteenth century Irish-born population resident in England and Wales. Surname data were derived from (1) the householders named in the index to Griffith's valuation of Ireland, a survey undertaken between 1846 and 1864, and (2) of Irish-born residents named in 1881 census of England and Wales. Visual representation of the Griffith's valuation isonymy matrix by multidimensional scaling (MDS) gives a result very close to the geographical distribution of Irish counties, and Mantel matrix correlation shows random isonymy between counties to be negatively associated with geographical distance, generally decaying according to a pattern of isolation-by-distance, with exceptions that can be explained in terms of Irish population history. Some 141,360 Irish-born residents in England and Wales at the 1881 census were assigned to an Irish county of origin, and random isonymy by county of birth for this group also shows a close correspondence to Irish geography. The Mantel matrix correlation between the Irish in Ireland and the Irish in England is 0.855, R(2) = 0.7306, indicating that the emigrant Irish in England were representative of the populations of the Irish counties from which they were derived. This result, together with the strong geographical patterning of surnames in Ireland, suggests that isonymy can be used to investigate the population structure and origin of Irish emigrant groups in Britain and potentially throughout the Irish diaspora. Am. J. Hum. Biol. 21:283-289, 2009. (C) 2008 Wiley-Liss, Inc.
