Adult survival and longevity in insects are key life-history traits, but their variation between sexes and individuals in natural populations is largely unexplored. Sexual divergence in senescence, the decline in survival with age is also poorly understood. Based on an intensive mark-recapture dataset of the butterfly Polyommatus daphnis, we aimed to assess whether adult survival is age-dependent, and to estimate life span distribution and abundance of males and females using Cormack-Jolly-Seber and Jolly-Seber models. Female survival slightly increased with date of emergence and slightly decreased with age, while male survival considerably declined with age. Mean life span of females (12.7 days) was ~50% higher than that of males (8.5 days), but two times higher if only the oldest 5% of each sex was considered (39 vs.19 days). Abundance of females (358 ± 14) and males (359 ± 11) was similar, but peak abundance of males preceded that of females by 11 days. Our results suggest that senescence is much more rapid in males than in females in this butterfly, which is in agreement with sexual selection theory. We also conclude that estimating life span distributions provides much more valuable information on the demography of natural populations than simply reporting the mean life span. In holometabolous insects, the adult stage is dedicated mostly to reproduction. This stage can be very short in many species which do not feed during this stage at all, but several others are able to acquire some nutrients to increase their fecundity and potential to find a mate, colonize new sites or migrate 1. Therefore adult survival, and consequently life span, may significantly affect reproductive success, and can be considered as important fitness components 2,3. There is a growing body of evidence on that senescence, the decline in performance and survival with advancing age 4 , is widespread in wild populations 5 , although most of the studies were conducted on vertebrates. Reproductive senescence means the decline in reproductive output, while actuarial senescence is the decline in survival with age 4. Theory predicts that sexual selection may produce sex-specific optima for traits that affect survival and ageing rate, often favouring a 'live fast, die young' strategy in males, i.e. higher mortality and more rapid ageing than in females 6,7. Under natural circumstances, estimating survival and ageing rate is very challenging, particularly for small-bodied, short-living insects 8 , thus insect demography is relatively understudied compared to that of vertebrates 9. There are only a handful of studies that properly estimated survival and ageing rate of both sexes in natural populations of insects. First Bonduriansky & Brassil 10 could detect ageing in males of an antler fly. In a wild population of neriid flies, Kawasaki et al. 11 observed higher mortality of males and no sign of ageing in females, while Dukas 12 provided evidence on senescence in honey bee foragers. Zajitschek et al. 13 also found higher mortality in males of a ...