A series of recent studies has shown an association between low serum cholesterol level and cancer, particularly of the large bowel. To explore this issue, serum cholesterol measurements were collected from 244 patients with adenomatous polyps of the colon, 182 patients with Dukes' A or B colon cancer, and 688 hospital controls who were diagnosed between 1979 and 1981 at a New York City hospital. The mean serum cholesterol levels were lower for patients with cancer (207.2 mg/dL) than for controls (219.5 mg/dL), with patients with Dukes' B cancer accounting for most of the difference. Patients with adenomatous polyps (219.8 mg/dL) were similar to controls. After adjustments for nutritional status using the serum albumin level, there were no statistically significant differences among any of the groups. We conclude that the low serum cholesterol level previously associated with malignancies, and colon cancer in particular, is a consequence rather than a cause of the cancer.