Uncertain paternity in primate quantitative genetic studies

Quantitative genetic studies in primates have generally been based on varying amounts of genealogical information. We consider the case where maternal relationships are known, but paternal relationships are only probabilistic (i.e., a limited number of males can be enumerated as equally likely sires for a given offspring). Using Henderson's [1988] average numerator relationship matrix method, we show for craniometric data from the Cayo Santiago macaque colony that heritability estimates are not greatly affected by the addition of incomplete paternal information. We then show through simulation studies that in order for there to be a substantial increase in power to detect significant heritabilities, the number of possible sires per offspring must be quite small. Given this restriction, we conclude that the current method of ignoring paternal relationships is probably adequate and that considerable effort would have to be expended in performing paternal exclusions before there would be a substantial increase in the precision of heritability estimates. © 1992 Wiley-Liss, Inc.