Conventional sampling theory has unfortunately ignored this problem; but a conventionally mathematical approach might not be of great help. Strategies of sampling based on the sociological characteristics of the population that interests the researcher are likely to be much more fruitful in the study of deviants. The deviant world one wants to study may be divided into several somewhat separate segments, such that various methods will be necessary for each; any study of homosexuality must reckon with this. Deviants frequently look for unconventional sources of medical, legal, and psychiatric services, either because they cannot afford what is conventionally available or, as is often true with regard to medical services, because they do not like the embarrassment and harassment they experience from conventional practitioners. Deviant activities tend to generate a special language, at the least to describe the esoteric events, people, and objects involved, and perhaps as a matter of symbolic differentiation from nondeviants as well.