Daily fluctuations of cortisol concentration in the blood or saliva have been repeatedly reported. However, several contradictions in the existing literature appear on this subject. The present study was performed to definitively establish options for testing adrenocortical function. To the best of our knowledge, this is the first study to evaluate parallel circadian rhythms in salivary and serum cortisol concentrations during a 24-h period. Twenty horses were examined under the same conditions. Blood and saliva samples were taken every 2 h for 24 h to determine the daily changes in cortisol concentrations of saliva and serum at rest and to determine the relationship between salivary and serum cortisol levels. Cosinor analysis of group mean data confirmed a significant circadian component for both serum and salivary cortisol concentrations (P < 0.001 in both cases). The serum cortisol circadian rhythm had an acrophase at 10:50 AM (95% CI, 10:00 AM-11:40 AM), a MESOR of 22.67 ng/mL, and an amplitude of 11.93 ng/mL. The salivary cortisol circadian rhythm had an acrophase at 10:00 AM (95% CI, 9:00 AM-11:00 AM), a MESOR of 0.52 ng/mL, and an amplitude of 0.12 ng/mL. We found a significant but weak association between salivary and serum cortisol concentrations; the Pearson correlation coefficient was 0.32 (P < 0.001). The use of salivary cortisol level as an indicator of hypothalamic-pituitary-adrenal axis activity may be warranted. However, the salivary cortisol levels are more likely to be correlated with free plasma cortisol than with the total plasma cortisol concentration.