Information on biological day-to-day variation is needed for detecting within-subject changes over time. In this study the daytime changes and the biological day-to-day variation of serum testosterone, follitropin, lutropin and oestradiol-17beta concentrations were investigated in 31 healthy males. To analyse daytime changes, blood specimens were taken at 0800 h, 1200 h, 1600 h and 2000 h during one day (n=31) and two days (n=8). The day-to-day variation was analysed from blood specimens collected at 0800 h on days 1 and 2 (n=31) and additionally on days 3, 4, 6 and 9 (n=8). The evaluation of the day-to-day variation was based on calculations of the within-subject (CVA+I) and between-subject (CV(G)) coefficients of variation. When the within-subject day-to-day variances were not too heterogeneous, they were used for the calculation of 95 % reference change limits. Serum testosterone and oestradiol-17beta concentrations showed a significant daytime variation; testosterone had higher serum concentrations at 0800 and 1200 h. A peak in the serum concentration of oestradiol-17beta occurred at 1200 h with a decrease towards the evening. There were no clear daytime changes in the serum concentrations of follitropin or lutropin. For different analytes the reference change limits were: serum testosterone +/- 32.0 %, serum follitropin +/- 24.1 % and serum oestradiol-17beta +/- 38.3 %. The reference change limit was not calculated for serum lutropin, as a high degree of heterogeneity and individuality was found. The interpretation of the results of hormone measurements requires recognition of the biological daytime and day-to-day changes of hormones. The reference change limits determine what changes are significant when monitoring the patient.