Objectives: Caregivers of persons with dementia can experience loss and grief long before the death of the person. Although the phenomenon of caregiver grief has been increasingly described, it is uncertain if caregiver grief has independent effects—separate from the well-studied construct of caregiver burden—on adverse outcomes such as caregiver depression. We sought to compare the effects of baseline grief and burden on caregiver depression at baseline and 2.5 years later. Design and Setting: A cohort study with 2.5 years of follow-up. Participants: Family caregivers of community-dwelling persons with dementia (n = 183). Measures: Participants completed questionnaires with scales that assessed caregiver grief, burden, and depression. Baseline grief and burden scores were included in a path analysis to predict depression at baseline and at 2.5 years. Results: At baseline, grief and burden had synergistic relationships with each other (P = .012), where the high levels of grief amplified the effect of burden on caregiver depression. Both grief and burden had longitudinal effects on caregiver depression at 2.5 years (P <.001 and P = .047, respectively), albeit with some differences in mechanism; both had effects which were indirectly mediated through baseline depression (P < .001 and P = .002, respectively), but only grief had a direct path toward depression at 2.5 years (P = .030), which was distinct from baseline depression. Conclusions and Implications: The findings highlight the need to identify and address caregiver grief in dementia services. They present a window of opportunity to improve caregiving outcomes, especially during the “latent phase” when caregivers encounter loss and grief but have yet to fully experience the debilitating effects of depression.