Two samples of patients with DSM-III-R anorexia nervosa, one identified by screening of a total population cohort and the other a mixed screening and referral group, were contrasted with a comparison group of sex-, age- and school-matched individuals. The rates and types of services consulted and treatment given were analyzed. The groups with anorexia nervosa tended to differ somewhat in these respects, although the numbers were too small for meaningful statistical calculations to be made. Only half of the total population sample had ever received any treatment for the eating disorder, even many years after anorexia nervosa onset in the teenage period. The implications of these findings are discussed both in terms of clinical needs and as they relate to previous research in anorexia nervosa, which has almost exclusively referred to non-population-based groups of patient.