We studied influences of dental care, food and storage on the reproducibility of salivary steroid levels. Cortisol (F), 17OH-progesterone (17OHP) and Progesterone (P) were measured using adapted commercial radioimmunoassays. Saliva samples of healthy adults (n = 15; m:8; f:7) were collected directly before and after dental care, and directly before and after breakfast with various foodstuffs. A second experiment investigated stability of steroids under different storage conditions. Four series of identical saliva portions (I: Native saliva; II: Centrifuged saliva; III: Saliva with trifluor acetate (TFA); IV: Saliva with 0.5% NaN(3)) were stored at room temperature and at 4 degrees C for up to three weeks. To demonstrate influences of repeated thawing and re-freezing of saliva on steroid values, saliva samples (n = 15) were divided into identical portions. These portions were frozen and re-thawed up to 5 times before measurement. Neither dental care nor intake of bread or milk effected the reproducibility of F, 170HP, and P. Steroid levels decreased significantly in the course of three weeks under different storage conditions (P < 0.001). This decrease was clinically relevant from the second week onward, with exception of NaN(3) treated samples. After repeated freezing and re-thawing 17OHP and P decreased slightly (about 5%). Only F decreased significantly after the third thawing (P < 0.001). The results show the usefulness of standardized handling of saliva samples for improving reproducibility and reliability of salivary steroid measurements.