The efficiency of different methods of the directed extension of the p
edigrees that are used to analyze the linkage of recessive genes was s
tudied. It was demonstrated that the addition of parents' relatives wi
th abnormal phenotypes into the pedigree allowed the total sample size
to be considerably decreased. All other methods of pedigree extension
are efficient only when they take into account the phenotypes of the
added relatives: at least one relative should have an abnormality. The
generally accepted notion that three-generation pedigrees are optimal
for linkage analysis was disproved for the case of recessive abnormal
traits. A choice between two alternatives for construction of pedigre
e samples was discussed: (1) the use of only nuclear pedigrees strictl
y tested for informative value with respect to linkage analysis and (2
) the introduction of any nuclear pedigree carrying an abnormality int
o the sample and the subsequent extension of the pedigree.