A search was conducted to find evidence of possible incestuous unions between the biologic parents of children involved in 2500 paternity cases. Suspicion was raised when either (1) a mother and her child possessed identical HLA phenotypes, or (2) the child appeared to be possibly homozygous for one maternal haplotype (i.e., one of the child's HLA haplotypes was a blank). These mother-child HLA-haplotype dualisms (MHDs) occurred in 5% of all cases. Frequency of exclusion of the accused men in cases demonstrating MHD, was compared with the remaining paternity cases. No significant difference was found in overall exclusion rates between MHD cases and controls when exclusion produced by HLA and red cell antigen systems were observed. However, there was a greater rate of exclusion in MHD cases when comparing exclusions produced by red cell antigen systems regardless of whether HLA tests excluded paternity (p less than 0.025). MHD cases involving teenaged mothers differed from control cases in frequency of exclusion of paternity only on the basis of red cell antigen phenotyping (p less than 0.005). The HLA system's usefulness in paternity testing is diminished when there is MHD; multiple, independently-inherited systems are relatively more useful in these circumstances. The search method detects only half of potential incest cases; proof of incest requires more extensive testing for homozygosity among other polymorphisms. Since calculations of likelihood of paternity are inappropriate in cases involving close consanguinity, detection and follow up studies are important. Data suggest that one-fifth of MHD cases may involve first degree consanguinity and that the incest rate among paternity cases may be as high as 2%.