Expressions are derived for the sample size required to achieve a given power in variance
component linkage analysis of a quantitative trait in unascertained samples. For simplicity an
additive model, comprising effects due to a single QTL, residual additive genetic factors, and
individual-specific random environmental variation, is considered. Equations are given relating
sample size to trait heritability for sibpairs, sib trios, nuclear families having two and three sibs,
and arbitrary relative pairs. The effects of nonzero residual additive genetic variance and parental
information are discussed, and a scale relationship for sample sizes with sibships and nuclear families
is derived. For larger sampling structures such as extended pedigrees the inheritance space is
randomly sampled and the relevant equations are solved numerically. Comparative power curves
are presented for sibships of size 2–4 and for an extended pedigree of 48 individuals. Simulation
results for sibpairs confirm the validity of the theoretical results.