ObjectivesTo identify depressive symptom trajectories and factors associated with trajectory group membership in the very old segment of the rapidly growing and long-living Mexican-American population. DesignLatent growth curve modeling was used to identify depressive symptom trajectories and multinomial logistic regression to identify factors associated with trajectory group membership. SettingData spanning three waves and 7years (2004-05, 2007-08, 2010-11) drawn from the Hispanic Established Populations for Epidemiologic Studies of the Elderly; homes of Mexican-origin elderly adults. ParticipantsCommunity-dwelling Mexican Americans aged 75 and older living in the southwestern United States (N=1,487). MeasurementsThe 20-item version of the Center for Epidemiologic Studies Depression Scale. ResultsThree trajectory groups were identified: low throughout, increasing, and high but decreasing. Activity of daily living disability was the strongest predictor of depressive symptoms, followed by social support. Foreign-born individuals were at greater risk than those who are U.S. born for high but decreasing depressive symptoms than for low depressive symptoms. ConclusionEarly detection and treatment of chronic disabling conditions, especially heart disease, cancer, visual impairment, and cognitive impairment, and increasing access to social participation should be the focus of treatment and intervention strategies for depression in very old Mexican Americans.