ABSTRACT Aims To describe rates of and reasons for culling and mortality of ewes between breeding and mid-lactation on New Zealand sheep farms; to investigate associations of these variables with farm demographic variables; and to describe rates of and reasons for culling of ewes at weaning. Methods Participants were a convenience sample of 34 farms from across New Zealand. Demographic data were initially collected for each farm via a questionnaire administered in-person to the flock owner or manager. During approximately 8 months from breeding to mid-lactation, ewe tally, culling and mortality data were collected and used to calculate various parameters related to flock performance and to investigate associations. During the main ewe-culling event at weaning, ewe-culling data were collected from 29/34 flocks participating in the study. Results There was considerable variation between flocks, but the between-flock mean replacement percentage was 29.2 (SD 5.0)%. Overall, a between-flock mean of 10.5 (SD 4.6)% of ewes presented for breeding were culled or dead/missing by mid-lactation and thus did not rear any lambs. Additionally, from 27 flocks that reported data on ewes’ success at rearing lambs, a between-flock mean of 3.9 (SD 2.5)% of ewes that remained alive at mid-lactation failed to rear any lambs, resulting in an overall between-flock mean loss of 23.1 (SD 6.3) potential lambs per 100 ewes. Two-thirds of ewe mortalities between breeding and mid-lactation occurred during the lambing period. Model results showed flocks with higher pregnancy scanning percentages had lower rates of culling and mortality between breeding and mid-lactation. However, apart from farm contour, from breeding to mid-lactation there were no associations for culling and mortality with farm size, flock size, number of ewes per labour unit, whether ewe hoggets (7–9 months of age) were presented for breeding, or duration of the breeding period. A between-flock mean of 16.5 (SD 8.3)% of ewes present at weaning were culled, and among mixed-age ewes, the most common reasons for culling at this time were age, incisor teeth defects and udder defects. Conclusions To reduce unnecessary ewe culling and mortality, attention should be focused on maximising conception rates, ensuring judicious culling decisions, optimising body condition score, and identifying farm-specific causes of death over the lambing period to facilitate targeted intervention strategies. Clinical relevance Identifying why and when ewes exit flocks, and comparing it with the data presented here, will facilitate the development of flock-specific interventions to reduce ewe culling and mortality. Abbreviations BCS: Body condition score; NI: North Island; SI: South Island