Experimental evolution offers biologists the "Beagle in a bottle", a testing ground for evolutionary hypotheses. These hypotheses are tested by tracking of the laboratory evolution of microorganisms. One aspect of evolution that has received scant attention, however, is the evolution of genome complexity: the dimensions and organization of the genome. Organisms have great diversity in their genome architecture and complexity, but we poorly understand the dynamics of genome-architecture evolution. Here we propose to use experimental evolution to address this topic in vivo for the first time. We will use Tobacco etch potyvirus (TEV), a plant RNA virus that has been developed as a model system for experimental evolution. We will study the evolutionary dynamics of both decreasing and increasing genome size and complexity by duplications of existing genes, complementing of viral genes by in trans expression, incorporation of new genes, changes in gene order, and segmentation of the viral genome. Subsequently, these viruses will be evolved, and we will determine consensus genome sequences, fitness and virulence. We expect that this research will help us to understand how the genome evolves after biologically-plausible, gross changes in architecture, and will help bridge the gap between experimental work and phylogenetic work. Hereby, we will shed light on important, fundamental questions in the evolution of complexity, which will be relevant to research in many related fields and inspire development of refined theories on the evolution of biological complexity. At the same time, our work will help to better understand how pathogens make important transitions in their evolution, which is relevant knowledge in a world threatened by an array emerging infectious diseases. Finally, our ultimate aim is to sustain and impel the broader dialogue on how biological complexity has arisen, and what the implications are for understanding our evolutionary roots as complex organisms.