In a population of about 1,300 Hanuman langurs (Presbytis entellus) living in one-male breeding structures at Jodhpur, Rajasthan, India, census data resulted in male:female sex ratios of 1.1 in infants, 1.5 in juveniles and 0.2 in adults. Long-term demographic monitoring of 144 known individuals up to the age of 30 months confirmed a higher juvenile female mortality. It is hypothesized that heavier maternal investment in males helps them to compete more successful. Once males emigrate from their natal troop, their mortality begins to exceed that of females. The strongly female-biased adult sex ratio seems to result from male-male competition: maturing males join male bands and are exposed to greater risks while traversing unfamiliar areas, occupying home ranges of harem troops and fighting with other males for possession of a female harem. A sample of 60 definitely known causes of deaths recorded over 21 years revealed that males died significantly more often than females in instantaneous ‘spectacular’ deaths from external causes, such as accidents, infanticide and predation, thus reflecting a high-risk, high-gain strategy. The sex difference in infanticide (18 males:9 females for 13 directly observed and 14 presumed cases) is not well understood because the commonly accepted hypothesis is that new harem residents try to shorten a mother’s lactational amenorrhea by killing her unweaned infant regardless of its sex. However, it might be that harem residents select against future competitors of their sons and at the same time spare female infants with which they can reproduce if their tenure (mean: 26.4 months) is long e