Behavioural changes before calving could potentially be used to predict the time of calving using advanced methods, to help improve monitoring by farm staff and identify those animals likely to experience dystocia. This study aims to identify whether there are differences in the behaviour before calving between heifers and cows, and between those which experience dystocia and those which do not. Twelve Holstein-Friesian heifers and twelve cows were selected. Half of each group had calved without assistance, and half had been assisted using a calving jack (for >1 minute). Their behaviour was recorded for 12 hours prior to the calf being expelled and for a 12-hour control period at a similar time of day, 1-4 days earlier. Continuous focal observations from video recordings were used to quantify the frequencies and durations of behaviours during 2-hour periods. For each 2-hour period, paired t-tests were used to examine differences in behaviour between the calving and control periods. Compared with the control, both groups of heifers showed a significant increase in the duration of tail-raising from 4 hours before calving with an average increase of 31min 51s ± 34min SD (t=3.11, df=10, p=0.011). This was earlier than observed in cows which showed a significant increase (of 47min 42s ± 19min 41s: t=8.04, df=10, p<0.001) in the final 2 hours. In unassisted groups, a significant increase in lying frequency compared with the control period started 6 hours before calving (heifers: p=0.048, cows: p=0.042) whereas in the assisted calving groups a significant change only occurred in the final 2 hours before calving. These results show differences between heifers and cows in their pre-calving behaviour which must be taken into account when predicting the time of calving from behaviour. However, for those assisted with a calving jack, no early-warning signs of a difficult calving were identified.