Although loneliness is often viewed as a serious problem of old age, few studies have addressed either its measurement or prevalence among older populations. The present study analyzed survey data from newspaper questionnaires circulated in three North American cities (N = 2,026) in order to examine the prevalence of loneliness across the life-span and some of its correlates in late life. Loneliness decreased across the adult life-span, with respondents age 65 and older the least lonely; elders were also more satisfied with their social relationships. Neither gender nor living alone was related to loneliness for older people. The data also partially confirm the desolation hypothesis, suggesting that desolation, or the loss of an intimate attachment, rather than isolation per se is a major correlate of loneliness in late life. Further, recency of loss was strongly related to increased loneliness. Directions for future research and intervention are discussed.