In a Nigerian town with a stable population of 20,000, a door-to-door survey was conducted, using a questionnaire involving a complete census and a simple neurological evaluation which had previously showed a 95% sensitivity and an 80% specificity for detecting neurological disease. Positive responders were evaluated and categorised, using agreed criteria for diagnoses. Nearly 100% cooperation was obtained. Life prevalence ratio for at least one episode of headache was 51/1000. Crude point prevalence ratio for migrainous headache was 5.3/100, and peak age-specific ratio was in the first decade. Prevalence ratio for epilepsy was 533/100,000 and peak age-specific prevalence ratio occurred in the 5-14 years age groups. The prevalence ratio for peripheral nerve disorders was 268/100,000, and age-specific prevalence ratio for tropical neuropathy increased with age. Prevalence ratio for stroke was rather low at 58/100,000, but was probably due to the people's attitude to the disabled elderly and high mortality of stroke which showed annual mortality rate of 70/100,000 which increased with age to 1519/100,000 per year in the eighth decade. Crude prevalence ratios (cases per 100,000) for others are 112 for neurological complications (including sciatica) of spondylosis, 15 each for poliomyelitis, motor neurone disease, development speech disorders, 10 each for syncope, hereditary neuropathies. Parkinson's disease, benign essential tremor, primary cerebellar degeneration, cerebral palsy, mental retardation, organic psychosis (probable intracranial tumor) and 5 each for muscular dystrophy, pyomyositis, spina bifida occulta, alcohol dependence and cerebral malaria. The implications of the findings are important for development of community neurological services in the developing countries.