The aim of this study was to establish a novel version of powerful “barcode viruses” as a tool for studying the replication and transmission dynamics of influenza virus in vitro and in vivo. Five barcoded APR8 viruses were firstly used to investigate infection kinetics (e.g. single- and multi-hit events, particle clumping and temporal aspects of co-infection) in vitro. This work demonstrated that the majority of infectious events in cell culture were single-hit events, but a significant number of infections were initiated by more than one virus particles (consistent with virus aggregation during release). Reassortment was found to occur efficiently and ubiquitously when near-isogenic viruses co-infected cells. The timing of asynchronous co-infection revealed that super-infection was possible if the second virus encountered the cell within 4 hr of the first virus. The super-infecting virus showed accelerated replication and enhanced yield, suggesting the second virus can take advantage of the already initiated replication machinery. Beyond this time point (coincident with the onset of progeny release from the first virus) the second virus was blocked by the initial infecting viruses. Five virus libraries carrying ~2000 individually identifiable variants were then generated for in vivo study. Amplification of the viral libraries in Madin-Darby canine kidney (MDCK) cells was achieved without substantial bottlenecking or preferential selection of specific sequences. Thirdly, two pilot studies in pigs demonstrated that intranasal inoculation resulted in substantial bottlenecking and a relatively small proportion of the inoculum gave rise to productive infection. Consequently, distinct viral populations were found in different nostrils and could persist over the course of the infection due to anatomical partitioning. Distinct sub-populations could be distinguished in other tissue sites (e.g. trachea and lung). Super-infection of individual pigs could occur around 2 days following primary exposure. The identity of the donor pigs could be determined by the barcode identities. In the first pilot study, around 600 variants were seen in each donor pig directly inoculated with approximately 6000 variants of the barcoded viruses. When a pig was co-housed with 3 donors, a typical transmission dose of 73-151 variants were seen. To further study the transmission dose between a single donor and recipient, a transmission dose defined as 30-60 on 2 days post contact (d.p.c) and 20-50 on 3 d.p.c was observed. To conclude, my PhD project has developed a powerful tool with a wide range of applications in influenza biology, particularly for studying transmission dynamics in a natural host system.