Background. Data on stability of plasma total cholesterol levels and its determinants among young adults are lacking. Knowledge of factors associated with low levels of plasma total cholesterol during young adulthood may help clarify the nature of associations between hypocholesterolemia and health or illness. Methods. Tracking of plasma total cholesterol was investigated using data from the baseline (1985-1986), Year 5 (1990-1991), and Year 7 (1992- 1993) examinations of the Coronary Artery Risk Development in Young Adults Study. Lifestyle (including dietary), physiological, medical, and psychological correlates of plasma total cholesterol were examined cross- sectionally at baseline using ANCOVA and multivariate logistic regression. The attributes of participants with persistently low plasma total cholesterol level after 7 years (i.e., remaining below the 10th percentile of sex- and race-specific distributions) were also examined. Results. The cohort in this analysis comprised 720 black men, 922 white men, 899 black women, and 944 white women who were between the ages of 18 and 30 years at baseline. Between 44 and 52% of those with plasma total cholesterol levels below the 10th percentile remained below the same percentile 7 years later. Among black men, a difference of 1 SD in age [3.7 years; odds ratio (OR) = 0.69; 95% CI = 0.52-0.91] and a difference of 1 SD in systolic blood pressure (10.5 mm Hg; OR = 0.73; 95% CI = 0.54-0.97) were independently associated with lower odds, respectively, of being in the lowest 10th percentile of the plasma total cholesterol distribution. Also among black men, current smoking and more calories from carbohydrates were associated with nonsignificantly higher odds of low total cholesterol level. Among white men, a 1 SD older age (3.4 years; OR = 0.78; 95% CI = 0.61-1.00) and a 1 SD higher physical fitness (118 see; OR = 1.41; 95% CI = 1.09-1.82) predicted lower and higher odds, respectively, of low plasma total cholesterol. Among black women, a 1 SD difference in albumin (0.3 g/dL; OR = 0.80; 95% CI = 0.63-1.03) was related to lower odds of low plasma total cholesterol. Among white women, the factors independently associated with low plasma total cholesterol were body mass index (OR for a difference in 4.0 kg/m2 = 0.73; 95% CI = 0.54-1.00) and γ-glutamyl transferase (OR for an increase in 9.6 IU/L = 0.41; 95% CI = 0.18-0.93). The independent predictive factors of stably low total cholesterol levels were age and uric acid among black men (both inversely related) and age, Framingham Type A Behavior (inversely), and calories from carbohydrates (positively related) among white men. Conclusion: Young adults with low plasma total cholesterol level have characteristics generally associated with good cardiovascular health. However, adverse attributes such as current cigarette smoking (notably among black men) may confound future associations between low total cholesterol and disease.