Ambulatory assessments of hypothalamus-pituitary-adrenal axis responses to acute natural stressors yield evidence on stress regulation with high ecological validity. Sampling of salivary cortisol is a standard technique in this field. In 21 healthy student teachers, we assessed cortisol responses to a demonstration lesson. On a control day, sampling was repeated at analogous times. Additionally, the cortisol awakening response (CAR) was assessed on both days. Participants were also exposed to a laboratory stressor, the Trier Social Stress Test, and rated their individual levels of chronic work stress. In pre-to-post-stress assessment, cortisol levels declined after the lesson. However, post-stress cortisol levels were significantly higher compared with those on the control day. Also, the Trier Social Stress Test yielded higher cortisol responses when using the control day as reference baseline. Associations between the CAR and chronic stress measures were observed solely on the control day. There were no significant associations between cortisol responses to the natural and laboratory stressors. Our results indicate that a control day might be an important complement in laboratory but especially in ambulatory stress research. Furthermore, associations between chronic stress measures and the CAR might be obscured by acute stress exposure. Finally, responses to the laboratory stressor do not seem to mirror natural stress responses. Copyright © 2012 John Wiley & Sons, Ltd.