BACKGROUND: We aimed to measure the overlap between the social networks of injection drug users (IDUs) and the patterns of related hepatitis C virus (HCV) infections among IDUs. METHODS: A cohort of 199 IDUs (138 of whom were HCV RNA positive) was recruited from a local drug scene in Melbourne, Australia, and was studied using social network analysis and molecular phylogenetic analysis of 2 regions of the HCV genome. RESULTS: Eighteen clusters of related infections involving 51 IDUs (37.0% of HCV RNA-positive IDUs) were detected; these clusters could be separated into 66 discrete pairs. Twelve (18.2%) of the 66 IDU pairs with related infections reported having previously injected drugs together; conversely, only 12 (3.8%) of the 313 pairs of HCV RNA-positive IDUs who were injection partners had strong molecular evidence of related infections. The social and genetic distances that separated IDUs with identical genotypes were weakly associated. Significant clusters of phylogenetically related sequences identified from core region analysis persisted in the analysis of the nonstructural 5a protein region. Genotyping and sequence analysis revealed 2 mixed-genotype infections. CONCLUSIONS: Static social network methods are likely to gather information about a minority of patterns of HCV transmission, because of the difficulty of determining historical infection pathways in an established social network of IDUs. Nevertheless, molecular epidemiological methods identified clusters of IDUs with related viruses and provided information about mixed-genotype infection status.