Document Type

Book Review


A population of German-speaking Mennonites in Mexico has accumulated a random inbreeding coefficient estimated as 0.0091 in four centuries despite numbers in thousands even at times of migration. The availability of two parental surnames per index adult permitted a more precise analysis than is usually possible except in Iberoamerican populations. Maternal names were more diverse than paternal names. This and some other phenomena were better described and quantfied by chi-square and the sign test than by the inbreeding coefficient. In particular, the analysis by chi-square showed the importance of extended families recruited for migration. Use of both maternal and paternal surnames to estimate random inbreeding corrects an upward bias that results from using paternal names only, especially when female names are more diverse.