To resolve conflicting findings on the association of complement allotypes with Alzheimer's disease (AD) we have studied the C4 phenotypes in 33 sporadic cases and in one family with familial AD. We found no association with complement alleles in familial or sporadic AD, even though a familial case had absence of the C4 null allele (C4BQ0). Our data do not suggest a role for complement genes in the pathogenesis of AD. It also seems that the C4B2 allele cannot be used as a marker for AD as has been suggested by others.