Background: in community studies the prevalence of depression is higher in women than men; however, in palliative care settings this relationship is usually less strong, absent or reversed. Aim: to identify reasons for excess depression among men receiving palliative care. Design: cross-sectional study. Setting/participants: we interviewed 300 patients recruited from a large hospice in South East London. Depression was measured using the Primary Care Evaluation of Mental Disorder. Results: the higher prevalence of depression among men was not explained by a higher prevalence of particular types of cancer nor confounding by other covariates. Possible effect modifiers were examined. Depending on others for help with basic tasks (eating, dressing, washing or using the toilet) was a risk factor for depression in men only, with 37.8% of dependent men being depressed compared to 2.4% of similarly affected women (OR = 24.3, 3.1–193.2, p = 0.003). We observed a dose-response effect between the level of dependency and depression in men ( p for trend = 0.01). Conclusion: depending on others for help with basic tasks appears to contribute to the burden of depression among men with terminal illness. This gender-specific association may explain why the usual gender differences in depression prevalence are not observed in palliative care.