In order to establish basic epidemiological data on chronic pain (duration > 3 months) in a rural population, a survey of pain symptoms was conducted by means of a postal questionnaire. The questionnaire was sent to a random sample (from the population register) of 15% of the population aged 25-74 (n = 1806) in two Swedish primary health care districts. The response rate was 90%. In a follow-up study individuals selected among the responders (neck-shoulder pain, widespread pain and controls without pain; n = 213) were examined and interviewed. They were requestioned about pain symptoms 24 months after the initial survey. Without sex differences 55% of the population had perceived persistent pain for 3 months and 49% for 6 months. Women experienced more multiple localizations of pain and had pain in neck, shoulder, arm and thigh to a greater extent than men. Prevalence of pain increased by age up to 50-59 years for both genders and then slowly decreased. The neck-shoulder area was the most common site of pain (women 32.9%, men 27.5%). Blue-collar workers and employers (including farmers) reported chronic pain to a greater extent than other groups. In 13% of the population, manifest pain problems were associated with reduced functional capacity. Examination of selected pain groups indicated a high proportion of unspecific musculoskeletal symptoms. Diagnosis with definite definitions, explaining the pains, were found in 40% of the individuals. Individuals with widespread pain had a higher pain intensity, more somatic symptoms, were more depressive and had the lowest scores for quality of life.(ABSTRACT TRUNCATED AT 250 WORDS)