The purpose of the study was to investigate whether a carbohydrate-rich versus fat-rich diet influenced the effect of training on the immune system. Ten untrained young men ingested a carbohydrate-rich diet [65 energy percent (E%) carbohydrate] and ten subjects a fat-rich diet (62E% fat) while endurance training was performed 3-4 times a week for 7 weeks. Maximal oxygen uptake increased by 11% in both groups. Blood samples for immune monitoring were collected before and at the end of the study. Blood samples were also collected, in parallel, from 20 age-matched subjects, and data from these subjects were used to eliminate day-to-day variation in the immunological tests. Independently of diet, training increased the percentage of CD3-CD16+ CD56+ natural killer (NK) cells from [mean (SEM)] 14 (1) % to 20 (3) % (P = 0.05), whereas the NK-cell activity, either unstimulated or stimulated with interleukin (IL)-2 or interferon (IFN)-alpha, did not change. Furthermore, training did not influence the percentages of CD3+, CD4+, CD8+, CD19+ or CD14+ cells. However, when the two diet groups were compared, it was found that the NK-cell activity had increased in the group on the carbohydrate-rich diet [from 16 (3)% to 27 (2)%] and decreased in the group on the fat-rich diet [from 26 (2)% to 20 (4)%] in response to training. The effect of training on unstimulated NK-cell activity was significantly different between the groups (P = 0.007). These data indicate that diet manipulation during training may influence natural immunity, and suggest that ingestion of a fat-rich diet during training is detrimental to the immune system compared to the effect of a carbohydrate-rich diet.