The apparent clustering of human cases of influenza A (H5N1) among blood relatives has been considered as evidence of genetic variation in susceptibility. We show that, by chance alone, a high proportion of clusters are expected to be limited to blood relatives when infection is a rare event.