In a spatially explicit simulation model of vegetation dynamics (‘VegeTate’), I labelled the initial mass of Pteridium aquilinum in each of 225 cells as a single, unique genet or clone. The physical environment was homogeneous and all genets shared the same phenotype. The aim was to discover whether and how the success of each genet was affected by its initial position relative to other genets and a competing grass species. In a scenario in which grazing generated a grass-bracken mosaic with complex spatial dynamics, the amount of growth of each genet ranged widely, from frequent extinction to mass increase by over 300 times. The main factor in the impact of position on genet growth was shown to be a benefit from the initial presence or proximity of a large mass of P. aquilinum. This was because a high density of P. aquilinum reduced local grazing intensity, allowing plant mass to accumulate and shifting the balance of competition in favour of P. aquilinum. Thus variations between cells in initial mass of P. aquilinum were greatly amplified. The implications of this amplification of initial differences between sites for population genetics are briefly discussed. Qualitative features of the spatial distribution of genets at the end of simulations matched reported observations on patchily distributed field populations of P. aquilinum. These features included dominance of a large population by a small minority of genets, widespread mixtures of a dominant genet and one or more subordinate genets, and the presence of patches of P. aquilinum formed both by agglomeration from neighbouring foci and by spread of dominant genets. Under less intense grazing, which allowed little or no development of vegetation mosaics, genet growth varied relatively little and initial variations in relative mass between genets were little changed. Based on this study, I hypothesize that any processes that generate non-linear spatial dynamics will also generate complex genet dynamics.