Data were collected between June 1997 and June 2000 by means of ethnographic mapping. Study participants were recruited in geographical areas known for crack use. In order to be eligible for the study, participants were required to be 18 years old or older and be current users of crack cocaine (defined as having used at least once in the week prior to the interview). Based on information obtained from the interviews, four types of crack users were identified. These types were differentiated by use frequency, protective factors (avoidance of harm related to drug use), access to crack, and life structure. "Stable users" were those who smoked crack occasionally and had many protective factors in place. Their daily lives were highly structured, and their access to crack was limited. "Tempted users" smoked crack more often than the "stable users," and they faced the possibility of chaotic use as a constant threat. All users in this group referred to temptations to spend more money on crack than they intended or could afford. They had their own connections with dealers, owned their own paraphernalia, and had developed their own use routines. "Grappling users" were characterized by more chaos in their lives than the "stable" and "tempted" users. They did not use crack regularly, but they had few protective mechanisms. They tended to be socially isolated with little external support to sustain strategies for avoiding crack use. Crack use tended to be an escape for this group. "Immersed users" were those who best fit the public stereotype of crack users. Crack use tended to dominate their lives, and they described crack as a drug that was impossible for them to control. They had difficulty in spending time, energy, or resources on anything other than crack. Since these findings reveal crack users as a heterogeneous group, programs that aim to address problems related to crack use should consider these individual differences in the development of intervention and treatment strategies. 1 table and 30 references