Yoh Iwasa reviewed the current state of sexual selection theory, and Andrew Pomiankowski and Rauno Alatalo demonstrated how the theoretical principles can be applied to specific empirical systems. In an excellent overview talk, Yoh Iwasa showed how the classical results regarding Fisher's runaway process of sexual selection can be put into a broader perspective. On the basis of an optimality principle, he demonstrated that, quite generally, a Fisherian 'runaway' does not occur when there is a cost of female choice. On the other hand, he made the important point that this and related results are based on equilibrium considerations. By means of a specific example, he showed that the runaway process easily leads to stable evolutionary cycles of exaggerated male traits and corresponding female preferences. In such a nonequilibrium context, the co-evolution of exaggerated traits and preferences can occur despite costs of female choice. In the second part of his talk, Iwasa discussed several variants of the 'good-genes' model of sexual selection. He derived the general result that costly male ornaments and corresponding female preferences can evolve when the cost for making the ornament decreases with male quality. Finally Iwasa pointed out some shortcomings of the current models of sexual selection, such as the lack of interest for e.g. sensory bias and direct benefits. This last point was addressed by Hanna Kokko, who analysed a model for a species with biparental care and mutual mate choice. She showed that there is a conflict between signals for genetic quality and parental quality, which may result in a 'runaway' in which the importance of parental care continually diminishes.