The current investigation reports findings on the temporal stability of happiness as assessed by the Memorial University of Newfoundland Scale of Happiness (MUNSH), and the stability of the predictor/happiness relationships in three subgroups of persons over 64 years of age. Respondents were interviewed twice, 18 months apart, on the MUNSH and on nine established correlates. Separate multiple regression analyses were used to assess the predictor/happiness relationship for each phase. The stability of happiness was evaluated by the inclusion of phase 1 MUNSH scores in the phase 2 predictor array. For urban and institutional persons the main independent predictors of happiness in both phases were housing satisfaction, health, activities, and changes in life events. For rural individuals only health and marital status remained consistent predictors for both phases. Happiness, greater in rural than in institutionalized persons, remained stable for all groups, with an average of 86% of the accounted MUNSH 2 variance due to MUNSH 1 scores. These results show that, although predictor effectiveness may differ across subgroups, happiness remains stable in later years.