A sequence of thirty-six nucleotides in the nsP3 gene of Ross River virus (RRV), coding for the amino acid sequence of HADTVSLDSTVS, was naturally duplicated during 1960-1979. This period coincided with the appearance of a new lineage of virus and a major outbreak of epidemic polyarthritis amongst residents of the Pacific Islands. This duplicated sequence has been found in all recent RRV isolates, but was present only as a single copy in the two preceding lineages of RRV which now are extinct.1 The objective of this study was to investigate the role of these 12 amino acids on RRV fitness and lineage replacement. We used reverse genetics to generate two RRV strains based on a natural human isolate of RRV named QML. One strain (QML1) mirrors the natural isolate with two copies of the element of interest in the nsP3 gene. The second strain (QML1-nsP3-del) only has one copy of that element and resembles those strains of RRV found in Australia prior to 1979. In vitro studies in Vero, BHK, PS-EK, human synovial cells and C6/36 cells showed no difference in growth kinetics and plaque morphology was similar in Vero, BHK and PS-EK cells. However in vitro competition studies demonstrated that the QML1 strain outcompeted the less fit QML1-nsP3-del strain in vertebrate cells (Vero, BHK, PS-EK and human synovial cells), but demonstrated no discernible difference in mosquito (C6/36) cells. This suggests that the duplicated 12 amino acid element may increase RRV fitness in vertebrate hosts but not in invertebrate vectors. In additional studies, a natural isolate of RRV (OREG) with the duplicated 12 amino acid sequence in the nsP3 gene has been shown to outcompete a second isolate from nature with one copy of the element (BIRD), in Vero cells and human synovial cells but not in other vertebrate cell lines (BHK or PS-EK) or invertebrate mosquito cells (C6/36). Investigations as to whether these in vitro observations apply in vivo to mosquito and mouse models are currently underway.