Four species of malaria parasites infect humans. Until recently, the focus has fallen largely on Plasmodium falciparum, which places a huge burden of disease across Africa. Attention is now also turning to Plasmodium vivax due to a growing awareness that it too can cause severe and fatal disease. P. vivax has the widest geographical spread, including Asia, the Middle East, Central and South America and the Western Pacific, totaling an estimated 70 million cases per year. Effective control strategies are urgently needed. Planners are hampered by the lack of information on the likely consequences of different interventions. Where data are lacking, mathematical models can be used to provide a 'best guess'. To provide valid predictions, the building blocks of the models must be correct. Basic measures of P. vivax infections are required, such as the duration of infections, relapse and clearance rates, the rate of new infections, infectivity to mosquitoes, and the relationship between clinical symptoms and parasite densities in the blood.P. vivax infection dynamics are difficult to study because people living in endemic areas may harbour several different infections at the same time, the density of parasites in the blood may transiently lie below the limit of detection, and because the parasites can lie dormant in liver cells for long periods and then relapse. Different infections can only be distinguished by high resolution genotyping which allows the characterization of the number of different infections and their diversity. Genotyping of samples from study participants taken repeatedly over time is underway for studies from Papua New Guinea. Statistical methods to analyse this data are not well developed. We propose to develop analyses, fitting models of the dynamics of infections to both the genotyping data and other sources, to estimate the above quantities. We expect that the estimates will be useful for both modelling and for P. vivax epidemiology.