This article treats aspects of family life for which there are comparable data from students at three or more points over the 75-year period 1924?1999, beginning with the 1924 Middletown high school surveys and drawing on the replications of 1977, 1989, and 1999. We reassess the Lynds' positioning of families in a trajectory of space and time that emphasized family decentering or dis-integration, and that assumed that technological changes extending a family's circulation in time and space reduced family solidarity. Although national data suggest that some family decline has indeed occurred since the 1970s, among Middletown students last surveyed in 1999 there is quite remarkable stability in several of the Lynds' indicators. Trends in topics of parent-child disagreement and students' perceptions of ideal parental traits suggest that family change does not necessarily mean family decline, and trends in emotional solidarity with parents, beginning in 1977, yield no evidence of decline. The Lynds assumed a linear, additive relation between family decline and nonfamily space/time utilization. We argue that their space/time indicators are problematic as correlates of family solidarity and suggest alternative measures. Students of 1999 reported spending less time with parents than did their predecessors, but their identification with and perceived closeness to parents was not lower than in earlier decades. At century's end, Middletown students seemed insulated from the widely reported national trend toward ?rapidly loosening family bonds,? and continuity rather than family decline seemed the dominant trend. As parental time with children is significantly related to, but accounts for little of the variance in, children's emotional solidarity with parents, further work is needed on predictors of emotional solidarity in families.