The seasonal variation of coronary heart disease mortality rates in New Zealand is analysed by age, sex and race using monthly national mortality data for the period 1970-83. A 35% variation from the winter peak to summer low is found in the crude mortality rate, but the size of the seasonal variation is age-dependent, being more pronounced in the elderly, and more so in males than in females. The hypothesis that respiratory infections are linked to coronary heart disease, and that their seasonal occurrence explains the seasonal variation in coronary rates, is examined by an analysis of the association between coronary disease and respiratory disease mortality rates. By partial correlation analysis and by examining the residual correlation after filtering the seasonal variation from both series, it is suggested that the season acts as a confounding factor to cause an apparent association between the two rates. After controlling for season there is a tenuous relationship, but it is apparent only in the elderly.