In a population survey in 1970/72 the prevalence of head nodules was 1.7% in 1098 cases of onchocerciasis in the rain-forest and 0.6% in 1128 cases in the Sudan-savanna of Cameroon. In a follow-up survey in the same villages 3-4 yr later more attention was given to the detection of head nodules, and the corresponding prevalences were 5.5 and 5.0%. In the follow-up survey a strong association was demonstrated between the presence of head nodules and lesions of the anterior and posterior segments of the eye. The relative risks of having eye lesions in patients with head nodules compared with those without were 2.9 and 7.5 in the rain-forest and savanna, respectively. In a group of 483 clinic cases with ocular onchocerciasis from the savanna palpable head nodules were detected in 23.6%, and in a further 140 selected cases from the same area with posterior segment eye lesions, head nodules were detected in 31.4%. Confirmation of the onchocercal origin of the nodules was obtained in doubtful cases by biopsy. Many head nodules probably remain undetected in onchocerciasis surveys. They are often very small, flat and hard, and tightly adherent to the underlying periosteum, and the patient often points out the presence of a nodule even when none was detected after careful examination. The presence of a head nodule is a recognized risk factor associated with a high prevalence of blindness in Central America, but there have been no detailed studies in African onchocerciasis. A trial is in progress to assess the effect of nodulectomy on the development of ocular lesions.