The rapid emergence of Acacia mangium as the key industrial plantation species in Indonesia has been followed by the equally rapid emergence of red root rot (Ganoderma philippii) as its potential nemesis. As a consequence, and on severely affected mineral soils in equatorial tropical environments in particular, A. mangium may no longer be capable of producing commercial yields after three rotations. In this experiment, 100-tree plots were established in commercial plantings of A. mangium at five sites which represented the range of growing conditions used for plantation establishment in Indonesia. Repeated monitoring at approximately 6-month intervals of above- and below-ground variables was used to explore relationships between measures of tree health and root rot. Crown colour and density were poor predictors of either the presence or severity of infections. Tree mortality increased more or less linearly with time and led to the progressive coalescence of initially discrete disease gaps. The average rate of disease development was about 0.3% per month, and average time from infection to tree death was conservatively estimated at around 1 year. Trees with more than four dead/missing neighbours had a >50% chance of being dead by the following monitoring. By the end of the monitoring period >40% of trees were classified as dead/missing, although this value varied from 20 to 70% depending on site, tree age and rotation. Canonical discriminant analysis correctly allocated >90% trees into their correct group on the first monitoring and similarly classified whether trees would be dead or missing in the following monitoring.