An epidemic of grapevine leafroll disease (GLD), caused by grapevine leafroll-associated virus 3 (GLRaV-3), was monitored over an 11-year period in Nuriootpa, South Australia. Inoculum originated from infected budwood, and initial GLD incidence at the time of transplanting in 1986 was 23.1%. Infected vines were planted in a random spatial pattern. Change in disease incidence was not observed until 8 years after planting, when disease incidence increased to 27.9%. Disease incidence increased to 51.9% by 1996. Disease progress and rate curves (dy/dt versus time) indicated that the logistic (R2 = 96.2) and Gompertz (R2 = 96.3) growth models would best describe disease progress. However, the logistic model, which has a simpler data transformation with fewer model assumptions, was chosen for the purpose of comparing this epidemic (South Australia) with a GLRaV-3 epidemic in Cabernet Sauvignon grapevines in New Zealand. The logistic rate of GLD spread with respect to time was 0.35 logit/year in South Australia and was nearly three times faster (1.19 logits/year) for GLRaV-3 spread in New Zealand. Ordinary runs analyses indicated that the arrangement of infected vines within rows in South Australia was random up to 8 years after transplanting but subsequently became highly aggregated. Thus, GLD-infected plants are contributing to new infections (i.e., there is evidence for plant-to-plant spread), and a biotic vector with a steep dispersal gradient from each point source is likely to be involved.