Abstract: Abstract Analysis of salivary cortisol is increasingly used to assess stress responses in horses. Since spontaneous or experimentally induced increases in cortisol concentrations are often relatively small for stress studies proper controls are needed. This requires an understanding of factors affecting salivary cortisol over longer times. In this study, we have analysed salivary cortisol concentration over 6 mo in horses (n = 94) differing in age, sex, reproductive state and housing. Salivary cortisol followed a diurnal rhythm with highest concentrations in the morning and a decrease throughout the day (P < 0.001). This rhythm was disrupted in individual groups on individual days; however, alterations remained within the range of diurnal changes. Comparison between months showed highest cortisol concentrations in December (P < 0.001). Cortisol concentrations increased in breeding stallions during the breeding season (P < 0.001). No differences in salivary cortisol concentrations between non-pregnant mares with and without a corpus luteum existed. In stallions, mean daily salivary cortisol and plasma testosterone concentration were weakly correlated (r = 0.251, P < 0.01). No differences in salivary cortisol between female and male young horses and no consistent differences between horses of different age existed. Group housing and individual stabling did not affect salivary cortisol. In conclusion, salivary cortisol concentrations in horses follow a diurnal rhythm and are increased in active breeding sires. Time of the day and reproductive state of the horses are thus important for experiments that include analysis of cortisol in saliva.