Twenty-three (AC)n repeat markers from chromosome 16 were typed in the parents of the 40 CEPH (Centre d'Etude du Polymorphisme Humain) families. Where parents were informative, the entire families were then typed. There were seven markers in which null alleles were demonstrated, as recognized by the apparent noninheritance, by a sib, of a parental allele. Four of these markers showed a null allele in a single sibship, while in the other three at least 30% of the CEPH sibships were shown to have a null allele segregating. One null allele was sequenced and shown to be the result of an 8-bp deletion occurring within the priming sequence for PCR amplification of the (AC)n repeats. In gene mapping or in application to diagnosis, the presence of a segregating null allele will not corrupt the linkage data but could result in loss of information. In isolated instances a segregating null allele may be interpreted as nonpaternity. The presence of a null allele may generate misleading data when individuals are haplotyped to determine the presence of linkage disequilibrium with a disease gene.