In high-yielding dairy herds, the difficulty associated with successfully detecting oestrus is a problem that limits the reproductive performance of the herd. Traditionally, the period when the cow stands to be mounted is taken as the oestrus period. Housing and management factors such as the use of concrete-floored loose-housing system and the fact that the synchronisation of ovulation is not commonly practised so that few cows are in oestrus simultaneously result in poor expression of standing oestrus. As recent studies have shown that the proportion of cows showing standing oestrus in these conditions is very low, it would appear that the commonly used behavioural characterisation and definition of oestrus is inadequate. The aim of this study was therefore to update knowledge on the behavioural aspects of oestrus in the modern dairy cows, to allow a new behavioural characterisation and definition of oestrus. Fifteen Holstein cows, in good health and kept in cubicle house, were studied during their second or third post-partum ovulation period. Milk progesterone profiles were used to predict the ovulation and oestrus period. Cows were observed using a video camera system. General behaviour, as well as instances of sexual, social and agonistic behaviour were recorded continuously from the beginning of the expected oestrus period until the day after oestrus had ended. The frequency of sexual behaviour initiated by the test cows (mounting, chin-resting and licking/sniffing the ano-genital region) and the time spent walking increase during the period of standing oestrus. Of these the most frequent were chin-resting (2–130 times per day) and sniffing/licking the ano-genital region (8–176 times per day). The number of standing mounts was very low (0–32), with only eight cows exhibiting standing mount once or more, during the expected period. Therefore, changes in the frequency of sexual behaviour initiated by cows appeared to be the best behavioural indicators of the oestrus period. Using this definition, 11 of the cows would have been considered to have been in oestrus, while four cows had silent ovulations. Standing to be mounted did not appear to be a sensitive enough measure to be used to define and detect oestrus in all cows, and some showed no visually detectable signs of oestrus in their second or third ovulation. We suggest instead that chin-resting and sniffing/licking of the ano-genital region of other cows should be used in the visual detection of oestrus.