A double antibody enzyme immunoassay was used to measure serum ferritin levels in several different control and tumor-bearing populations collected from two institutions. The control groups consisted of normal volunteers, smokers, and Latter Day Saints. No statistically significant differences were noted in ferritin levels between pairs of these groups. Differences were noted among the normal groups when separated on the basis of age and sex, with higher ferritin levels in individuals older than 32 years of age and in men. By one-way analysis of variance, most control groups and subgroups were shown to have significantly lower levels (P less than 0.05) than the head and neck cancer group, with the exception of male smokers, who had levels comparable to male head and neck cancer patients. Serum ferritin levels were higher in head and neck cancer patients than in controls; however, there was no difference when compared with patients with other types of solid malignancies or when considering the anatomic site of the head and neck lesion. Ferritin levels were significantly (P less than 0.05) higher in patients with advanced (Stages III and IV) cancer than in those individuals with Stage I or II cancer. In patients with no evidence of clinical disease 5 years after treatment, the ferritin level had essentially returned to normal. In a group of head and neck cancer patients followed longitudinally, a significant decline in ferritin levels (P less than 0.05) was seen by 5 months after the completion of successful treatment. Furthermore, ferritin levels showed a tendency to increase or remain at high levels in patients with a poor prognosis and to decrease in those patients with a favorable prognosis, giving support to the contention that ferritin may prove to be a valuable tumor marker in head and neck cancer.