This paper reports the results of an experiment to determine whether sex ratio has any effect on the spawning efficiency of common dentex, Dentex dentex, broodstock. The paper also reports preliminary data on diurnal changes in water concentrations of steroids (measured on two separate occasions) in the same groups of fish. These latter observations were made to determine whether, using a non-invasive procedure, it was possible to relate hormonal changes in the fish to the "success of spawning." Mature fish were placed in broodstock tanks at three different sex ratios (Group A, 4 female:1 male; Group B, 1 female:1 male; and Group C, 1 female:4 male) and regularly monitored for egg production. At 18 days after the start of spawning, water samples were removed from all three tanks at 130 min intervals over a period of 24 h. At the end of spawning, water samples were removed at 4 h intervals. Free and conjugated steroids were extracted from the water and measured by radioimmunoassay. The highest number of eggs released, percentage of viable eggs, number of spawnings, spawning index, and relative fecundity were observed in Group B (1 female:1 male). The sex steroids testosterone (T), 11-ketotestosterone (11-KT), 17,20beta-P, and 17,20beta, 21-trihydroxy-pregn-4-en-3-one could be readily detected in water from all tanks. Concentrations of the conjugated forms (sulphate and glucuronide) were higher than those of the free form--with the exception of sulphated T and 11-KT. The concentrations of glucuronidated 17,20beta-P, sulphated 17,20beta-P, and free T appeared to be associated with the number of females in the tank and glucuronidated 11-KT concentrations with the number of males. In the first set of water samples (taken during the spawning period) there was also evidence for a cycle in steroid release with an acrophase during the afternoon and early evening, a few hours before the actual act of spawning. We were not able to show any association between steroid concentrations and success of spawning. However, since the data were derived from a "snapshot" of just a single day in what was a prolonged spawning period, this was perhaps not surprising. The results from the present study will help us to design experiments that are more able to answer this question.