Summary Female ungulates are often selectively harvested according to their reproductive status. Because ungulate population growth depends heavily on adult female survival, it is crucial to understand the effects of this selective harvest. Recent studies revealed persistent individual differences in female reproductive potential, with a positive correlation of reproductive success over consecutive years. If current reproduction is correlated with lifetime reproductive success, then selective harvest of non-lactating females should remove individuals of low reproductive potential, with lower impact on population growth than random harvest. If lifetime reproductive success has a genetic basis, selective harvest may also increase the proportion of successful females. We used an individual-based model to understand the short-term effects of harvest intensity and hunter selectivity on population dynamics, accounting for both heterogeneity in reproductive potential and orphan survival. We also explored the long-term effect of harvest as a selective pressure on female heterogeneity. Selective harvest of non-lactating females reduced survival to primiparity compared to random harvest, because of high harvest rates of pre-reproductive females. After primiparity, however, females of higher reproductive potential had higher survival under selective than random harvest. Therefore, the overall effect on population dynamics depends on a trade-off between a high harvest of pre-reproductive females and a reduced harvest of reproductive females with high reproductive potential. Female heterogeneity and the length of the pre-reproductive period affected this trade-off. Over the short term, high heterogeneity in reproductive potential of pre-reproductive females made selective harvest the most effective strategy to maintain a high population growth rate. With low heterogeneity and little effects of orphaning on juvenile mortality, however, random harvest had a lower impact on population growth than selective harvest. Over the long-term, selective female harvest may increase the proportion of successful reproducers in the population. Synthesis and applications. Selective harvests of non-lactating females appear justified only if female heterogeneity in reproductive potential and/or orphan mortality are very high. Because pre-reproductive females will be subject to intense harvest, selective harvest may reduce population growth rate compared to random harvest in species with late primiparity, especially if most pre-reproductive female normally survive to primiparity. When heterogeneity in reproductive potential and orphan mortality are low, random female harvest appears preferable to selective harvest. This article is protected by copyright. All rights reserved.