Survival rates are crucial demographic parameters for modelling the population dynamics of the Cape Griffon vulture. Previous survival estimates based on ring recoveries are seriously flawed and have been heavily criticized. To provide better estimates, 97 nestlings were individually colour-ringed in the winterrainfall region of South Africa from 1979 to 1987 and an intensive resighting campaign was mounted for 17.5 years (1980 to mid-1997). Although survival rate is known to vary with age, particularly in the first few years of life, this study indicates that it is also a function of calendar year. Supplementary food was provided for the study population from 1984 onwards and there was a concomitant increase in the survival rate of first-year birds. Eighty generalized linear models were built to estimate survival rate as a function of age, calendar year, cohort and supplementary feeding. First-year survival rate was estimated as 42.3% per annum (95% limits: 29–56%), increasing to 68.9% (54–81%) after the introduction of supplementary feeding. Age-specific survival rates were 88.8% per annum (77–95%) in the second year, 78.7% (64–88%) in the third year and 67.6% (58–76%) for birds aged four years or more. Our survival estimates are higher than previous values and have greater reliability because they are based on direct observations of known-age birds. The loss of colourrings is a serious impediment to estimating survival among older birds and the technique described here was not used to estimate adult survival.