The hypothesis that certain milk products contain a cholesterol-lowering "milk factor" was tested in adolescent schoolboys whose diets were complemented with 2 L of skim milk, yoghurt, or full cream milk daily for 3 wk. After a fall in all serum lipids during the precomplementation wk, serum total cholesterol and low-density lipoprotein cholesterol continued to fall on skim milk; in contrast, it rose for the first 2 wk on yoghurt or full cream milk. These changes correlated with dietary fat and cholesterol intakes. Total cholesterol returned to base-line values during the 3rd wk on yoghurt or full cream milk. High-density lipoprotein cholesterol and percentage high-density lipoprotein/total cholesterol rose transiently in all three groups, with the highest levels being recorded on full cream milk. Serum triglycerides tended to decrease in all groups. No convincing evidence of a milk factor could be found, but skim milk appeared to have a cholesterol-lowering effect at least partly due to its low lipid content.