The accurate detection of oestrus in beef and dairy cows prior to AI or embryo transfer (ET) is essential if optimal pregnancy rates are to be acheived. In dairy cows under normal husbandry conditions only 55% of normally cycling animals are correctly detected in oestrus and inseminated at the appropriate time (Esslemont and Bryant 1976). The use of pedometers, changes in milk yield, milk progesterone and trained sniffer dogs are all methods which can significantly improve the oestrus detection and hence pregnancy rates acheived by AI (Phillips and Schofield, 1988: McLeod, Foulkes, Williams and Weller, 1991; Jezierski, 1988) . However, oestrus detection by visual monitoring is still the most widely used method of detection.

To ensure high pregnancy rates from AI or ET when visual oestrus detection methods are used the form and range of behaviours which cattle perform needs to be understood. There are a wide range of oestrus behaviour patterns displayed both within and between individual cows so deciding on the correct time for AI or ET depends on interpreting the available evidence. This paper reports the behaviour of maiden heifers during natural oestrus and following a pharmacologically induced oestrus (designed to synchronise the time of oestrus for a number of heifers). By understanding the range of oestrus behaviour patterns guidelines can be established for observation methods and the interpretation of oestrus behaviour. This should enable the prediction of the optimum times for AI or ET. A corresponding improvement in the number of pregnancies should be acheived.