Bird song is known as the acoustic analogue of the peacock’s tail, and is often used as a model in sexual selection studies. Most research has focused on between-individual differences in song traits and investigated questions about how aspects of male quality can be signalled and how these can be translated into fitness benefits. However, bird song is a very special in a sense that it can also show a considerable variation within individuals, while spatio-temporal patters define biologically important patterns of variations at the between-population level. My research group has been studying the song of the collared flycatcher (Ficedula albicollis), which has a modestly complex system with males using 15-50 different syllable types in their individual-specific repertoires that they vary with high plasticity. In our earlier studies, by adopting the classical focus from behavioural ecology on between-individual variations, we determined the information content of male repertoires its relationship with mating success. More recently, we investigate how songs change within individuals across different temporal windows (within- and between days, and between years) and how such reaction norms differ among individuals and what are the fitness consequences of this variation. At a higher level, we study compositional changes in song at the between-population level, which open horizons for understanding the dynamics of cultural evolution. In my talk, I will provide examples from our model species to demonstrate the evolutionary relevance of the complex hierarchical organization of bird song. [more]