Two hundred twelve bereaved elders rated marital adjustment using items drawn from the Locke and Wallace (1959) Marital Adjustment Test and completed the Beck Depression Inventory 2 months, 12 months, and 30 months after the loss of their spouses. Their responses were compared with those of 162 nonbereaved individuals of comparable age who were tested at the same times. More positive ratings of marital adjustment were made by bereaved subjects than by nonbereaved subjects. Among nonbereaved elders, more severe ratings of depression were associated with lower ratings of marital adjustment. In the bereaved sample, however, the opposite was found: More severe ratings of depression were associated with higher ratings of marital adjustment. This pattern of results changed only slightly over the 2.5-year course of bereavement and was not influenced by gender. These results are discussed in terms of cognitive processes (e.g., idealization) that influence retrospective assessments of marital adjustment during bereavement.