During the post-Civil War decades, some of the homeless went "on the road," while others gravitated to the cities. There was considerable overlap between the two groups, but those who traveled in search of work (or, sometimes, adventure) were generally younger than those who remained permanently in one locale. This dual aspect would continue to define homelessness until the 1940s, when the effects of war and structural changes in the economy led to a sharp decline in the number of persons riding the rails. After 1945, homelessness would undergo a drastic change as an aging population of destitute men became confined, for the most part, to the deteriorating skid row areas of cities. Homelessness, which in the 1930s had reemerged as an important national issue, now reverted to what it had been before the Civil War-a strictly urban problem. (Conquergood , 2001) Even in the cities, the homeless became largely invisible to all but the police. The lack of concern for this impoverished group made the skid rows ripe for urban renewal, and in the 1960s and 1970s most of the old lodging house districts in American cities were demolished.
The level of significance we ascribe to homelessness very much depends on how the term is defined. In conducting the first census of the homeless in 1933, sociologist Nels Anderson identified a homeless person as "a destitute man, woman or youth, either a resident in the community or a transient, who is without domicile at the time of enumeration. Such a person may have a home in another community, or relatives in the local community, but is for the time detached and will not or cannot return." This succinct definition recognized that a homeless person could be either a permanent resident of a community or a traveler, that the condition of homelessness could either be voluntary or involuntary, and that family relationships were significant in determining whether or not a person became homeless. 1 All of these aspects are important for understanding the phenomenon historically. (McNally, 1999)
Counting only those "without domicile," however, implies that only persons literally without a roof over their head, or forced to sleep in public or private shelters, are genuinely homeless. Such a restrictive definition seriously underestimates the level of homelessness in society. People sleeping outdoors are difficult to count, and even diligent investigators will miss many, as census enumerators discovered in 1990. Anderson's definition also sidesteps the fact that homelessness is often a transitory condition. A person can be temporarily domiciled at one point yet still be functionally homeless. Recent studies have shown that many persons living on the street or sleeping in shelters are able, from time to time, to find accommodations with family or friends. (Millstein, 1998) These arrangements are almost always temporary, however, and in most cases such individuals are back on the street in a relatively short time. The best contemporary estimates indicate that for every person in a shelter or on the street on a given night, three or four times as many have been homeless at