Today the great majority of noninstitutionalized elderly widows live alone, a striking increase from a quarter-century ago. A noticeable difference has occurred, however, in trends by age; the proportion of the young-old widows living alone is starting to decline. while that of the old-old continues to increase. We use a model suggested by earlier studies to explain the emergence of this difference, and assess the prospects of its continuing over the next three decades. We find that the recent differential change in the proportions of younger and older widows living alone is due primarily to a differential change in kin availability that has emerged as the baby boomers’ parents have begun to reach retirement age. Over the next decade, the same type of differential change by age in kin availability will continue; living alone is likely to become less common among young-old than among old-old widows, in a reversal of the pattern of the last quarter-century. In the first two decades of the next century, as the baby boom affects kin availability among the old-old, and as the subsequent baby bust affects that among the young-old, the age pattern of living arrangements among elderly widows will reverse once again.