This study examines the independent and interactive effects of perceived social support (positive ties) and relational stress (negative ties) on loneliness in later life. To that end, we differentiate the sources of social support, tap multiple items to measure relational stress, and analyze a large cross-national dataset containing probability samples of older adults across 28 countries ( N = 12,449). Our analysis demonstrates that the relationship between perceived social support and loneliness is source specific. A key finding is that perceived social support particularly from family is negatively associated with loneliness. Older adults who name friends and ‘others’ (e.g., coworkers, neighbors, acquaintances) as the most important source of support exhibit higher levels of loneliness than those who choose kin members. By contrast, relational stress, regardless of its source, is positively associated with loneliness. Moreover, this positive relationship between relational stress and loneliness is stronger for those who select others as the primary source of social support, lending credence to the concept of reverse buffering. Taken together, our findings underscore the importance of the sources of social positivity and social negativity in examining the interrelationships between positive ties, negative ties, and loneliness in later life.