Objective: To evaluate the impact on nutrition-and health-related practice of two methods of delivery of a nutrition and health intervention in Irish full-day-care pre-schools: training of pre-school managers only or training of managers and their staff. Design: A simple randomised study with pre-schools divided into two training groups: 'manager trained' and 'manager and staff trained'. Direct observational data -food and fluid provision, physical activity, outdoor time, staff practices and availability of nutrition and health resources -were recorded during one full day spent in each pre-school both pre-and post-intervention, using a specifically developed and validated Pre-school Health Promotion Activity Scored Evaluation Form. Post-intervention, self-assessment data were also collected using the same evaluation tool. Setting: Pre-schools, Midlands of Ireland. Subjects: A convenience sample of forty-two pre-schools registered with the Irish Health Service Executive. Results: From pre-to post-intervention, significant improvement (P , 0?05) in nutrition-and health-related practice was observed within both intervention delivery groups in all areas evaluated: environment, food service, meals and snacks. No additional effect attributable to staff training was observed. Scores assigned by direct independent observation were lower than pre-school self-assessment scores. Conclusions: The implementation of a training intervention in pre-schools significantly improved practice with no significant benefit of additional staff training. Direct independent observation is required to quantify practice accurately.