Our results clearly show that repertoire size and composition are highly repeatable in our population, suggesting on the one hand that great tits may sing their complete repertoire at dawn, and on the other hand, that our recording methods appeared to be highly reliable. Although we cannot exclude differences at population level, the methodological differences may explain why our repertoire size was larger, and it may also suggest that repertoire size may have been underestimated in the previous studies. This may explain the reported changes in repertoire composition as well as the repertoire flexibility. The potential problem of underestimating repertoire size has been recognized previously as a possible confounding factor when studying great tit song. British researchers have already mentioned that they may have spent more time recording individuals with low song production, and that as a consequence, large LY2835219 repertoires may have been underestimated. Moreover, Franco & Slabbekoorn also considered the possibility that they had missed part of the repertoire, since they mentioned that they detected on average 90% of the song types per individual, and when they could only record the minimum number of phrases they were only able to detect 80% of an individual’s repertoire. As a consequence, if repertoires are underestimated during the first phase of a study on repertoire plasticity or learning, the ability to recognize ‘new’ songs in the later phases may be confounded by the fact that a part of the repertoire, although present, was not previously recorded. On the other hand, the time at which recordings were collected, which seems to differ from the one used by McGregor & Krebs, may also have played a crucial role in the assessment of repertoire. From the description in their article, it seems that recordings were obtained during daytime singing, while we only recorded individuals during dawn chorus. Differences in singing activity with time of the day or context have been already established for other species. For instance, common nightingales differ in diurnal and nocturnal singing activity and the banded wren sings longer during dawn chorus and preferentially uses specific song types in territorial interactions. Moreover, some evidence indicates that great tit males only use particular song types in territorial encounters, and males prefer to match the song of the neighbours. If great tits only use particular song types during daytime singing due to the territorial interactions, it would be possible to miss part of the repertoire while making only daytime recordings, which could also lead to an underestimation of the repertoire. Context dependent use of song repertoires could also have an effect on repertoire assessment during dawn chorus recordings. If great tit males use specific song types depending on the context, there is still a possibility that some song types are not used during dawn chorus.