This paper reports the main conclusions of a series of experiments concluded at Adiopodoume in the forest zone of the Ivory Coast between 1980 and 1990 on the temporal spread of African cassava mosaic virus (ACMV). The experiments sought to gain a better understanding of the epidemiology of the disease it causes and to facilitate the assessment of control measure. The course of ACMV epidemics over time was shown to depend closely on crop age and planting date. These relationships were expressed mathematically and ACMV progress curves were modeled combining a direct interaction between an overall exponential decreasing susceptibility to infection with crop age and a sinusoidal temperature-driven seasonal fluctuation in amount of spread from outside sources. After being validated using 1930s data from Kiwanda in Tanzania, this model was extended to incorporate host plant resistance, spread within plantings and yield losses. Simulation studies showed that when reversion (non-systemicity of the virus) does not occur and when cuttings are not selected preferentially from healthy plants, disease incidence increased in successive plantings of the same clonal stock and ultimately reached 100%. This occurred whatever the degree of host resistance, albeit after different periods. By contrast, with reversion and/or cutting selection, disease incidence may reach equilibrium values below 100% in resistant cultivars. At such equilibria, the effects of reversion and/or cutting selection balance the new virus transmissions by whiteflies. This emphasizes the potential of resistant cultivars to control ACMV by exploiting their ability to revert, as such cultivars not only suffer yield loss when infected, but are less likely to become heavily infected, even after many cycles of crop production.