A modelling approach was developed to identify the most likely composition of life-cycle phenotypes and mortality rates for Calanoides acutus found within the Scotia Sea. A stage- and age-structured model was parameterised with independent measurements of early copepodite stage duration, growth rate during the productive period and mortality and weight loss during the over-wintering period. Four possible life-phenotypes were simulated, varying according to whether (1) they spent their first winter as a CIV or CV stage, and (2) they persisted for between one and several years. The mortality rate coefficient during the epipelagic period was allowed to vary between 0.01 and 0.1 d(-1). Model runs determined the effect of all possible combinations of these variables on the abundance of the late copepodite stages, and predictions were compared to a comprehensive empirical data set, encompassing observations made over the last 60 yr. Best fits were achieved when 70% of the Scotia Sea population overwintered as a CV stage and 30% as CIV, with the majority of individuals reaching adulthood in their first year (after 1 overwintering period). During the epipelagic period, the population experienced mortality rate coefficients close to 0.06 d(-1). This high rate makes it unlikely that the earliest recruits survive to the end of the productive season. Predation was responsible for over 80% of mortality during winter. Starvation, which made up the remainder, mostly affected individuals that were spawned late in the season. The total rate of overwintering mortality was less than 0.007 d(-1). Sensitivity analyses demonstrated that model predictions were robust and identified the most likely causes of interannual and spatial variation in the observed data sets.