Seasonal variation of plasma cholesterol levels was studied in 1446 hypercholesterolemic 35- to 59-year-old male participants in the Lipid Research Clinics Coronary Primary Prevention Trial placebo group. Each man's serial cholesterol data, obtained at bimonthly intervals for 2.0 to 6.5 years, were analyzed as a separate periodic time series, and distributions of cycle zeniths and amplitudes were calculated. A highly significant (chi 2= 706, 2 degrees of freedom) seasonal effect, 7.4 mg/dl higher on December 30 than on June 30, was found. This effect was similar among the 12 LRC centers, including such disparate climates as those of Minneapolis and San Diego, and tended to be larger in the southern centers. Its magnitude was independent of baseline levels of plasma cholesterol and other baseline characteristics. Observed seasonal differences in weight and diet explained less than one-third of the seasonal variation in plasma cholesterol levels. Plasma low- and high-density lipoprotein cholesterol levels, analyzed similarly, also showed significant synchronous seasonal cycles. Plasma triglyceride levels showed a weaker irregular seasonal pattern, highest in midsummer and late autumn and lowest in spring. The etiologies and mechanisms for these seasonal patterns remain largely unknown.