An extensive table shows estimates of the prevalence of alcohol problems among contemporary American homeless populations from studies conducted since 1980. Two other tables provide similar information for homeless drug abusers and mentally ill persons. Other tables provide recent prevalence estimates of ADM problems among homeless populations by region, sampling site, sex composition of the sample, assessment method, and by those sampled in shelters using accurate assessment methods. In comparing the data by sample site, composition, and assessment method, the study concludes that no uniform profile of prevalence and characteristics of the homeless is possible given the diversity of study definitions of the homeless population in general and specific subgroups within it. The choice of sampling site also greatly affects the estimation of prevalence rates because of substantial underlying differences in clientele. The article recommends that future studies in this area clearly explain definitions, sampling methods, and measurement techniques. Also, they should analyze data so as to reflect the heterogeneity of the homeless population. 8 tables, 156 references.