Ten studies in the literature concerning regression of rabbit atheroma were re-examined. In studies where cholesterol content was referred to weight, a degree of regression was noted in 3/4 studies. Such regression might at least partly have resulted from the dilution effect of the atheroma contents when results were expressed on a weight basis. By contrast, when results were referred to length or protein, partial regression was seen in only 1/4 studies. Mild atheroma induced by short-term cholesterol feeding did seem to regress in 2/2 studies.