Abstract: (1) Sex and age data were collected on 60 111 feral horses (Equus caballus L.) removed from eighty-nine areas in Nevada, Wyoming, and Oregon between 1976 and 1987. (2) Sex ratios of young seldom differed from parity; however, sex ratios of adults were commonly skewed toward females. No evidence of differential capture probability between adult males and females could be detected; therefore, skewed adult sex ratios were attributed to differential survival. (3) Age-specific trends in sex ratios indicated that the proportion of males steadily decreased from near parity in foals, to lows of 0.61-0.77 in the 4-5-year age-classes. The trend then reversed with males becoming predominant (1.08-1.36) in the > 10 years age-class. (4) Population simulations suggest that survival diffentials of 0.05-0.07, favouring females to 4 years of age, and 0.02-0.04 favouring males in older age-classes were required to mimic observed age-specific sex ratio changes. To obtain the high proportion of males in the > 10-years age-class, onset of senescence also had to be earlier for females. (5) Causes for differential survival in the immature age-classes are uncertain, but may relate to behavioural or metabolic differences between the sexes. Differential survival between adult males and females is attributed to differences in the energetic costs of reproduction and disparity in their reproductive life spans.