The inbreeding coefficient (F) and its random (Fr) and non- random (Fn) components are general measures of genetic structure of populations and can be estimated by marital isonymy. However, in many situations the small number of cases of marital isonymy and occurrences of pseudo-marital isonymy (adoption of husbands’ surnames by wives before their marriages) affect estimates of F and Fn from surnames. The repetition of pairs of surnames in married couples (RP) avoids these two problems. RP in one-surname “lineages” taken one at a time or averaged, equals four times the kinship coefficient (Φ or k) and twice the coefficient of relationship (Ri) of those marrying into each lineage. Therefore RP and its random and nonrandom components in one-surname “lineages” share the universality of inbreeding coefficients as measures of population structure, but reduce the sample-size problem and are not subject to the problem of pseudo-isonymous marriages. Among marriages in England and Wales in 1975, the weighted mean RP of 5,642 “lineages” of rare British surnames is .001196 which is 24% above the random component (RPr) of .000964. The weighted mean of 610 moderately rare British surnames yields weighted mean RP of .001064,10.4% above random. Six common English surnames yield weighted mean RP of .001051, 9.0% above random. Three Welsh surnames yield weighted mean RP of .001961, 103% above random. Two Irish surnames, .000778,19% below random; and a small sample of one Scottish surname gives RP = .00714 which is 26% below random. Other recent immigrant one-surname RP values greatly increase the mean and variance but one can simplify interpretation by excluding them.
Lasker, Gabriel W.
"Repeated Surnames in Those Marrying into British One-Surname “Lineages”: An Approach to the Evaluation of Population Structure through Analysis of the Surnames in Marriages,"
1, Article 4.
Available at: https://digitalcommons.wayne.edu/humbiol/vol60/iss1/4