Abstract: This study investigated drinking behaviour of unhandled, semi-feral Dartmoor ponies. Aspects studied were drinking behaviour, latency to drink from novel unfamiliar drinkers after transport, preferences for different types of artificial water drinkers, effects of mixing with unfamiliar ponies and group size, on drinking behaviour, and the effect of a simulated market on the latency to drink. Ponies were tested in groups of three or six animals, or as individuals in test pens that were equipped with three water drinkers: bucket, automatic drinking bowl, flowing water trough. Behaviour was recorded using time-lapse video. An individual pony drank on average 10 l per day. Ponies also drank, but at a lower rate, during the night. The latencies to drink after 4.5 h of transport showed large variation, but most ponies drank within the first hour after being transported (all groups 80.5 ± 32.94 min, mean ± SEM). In the individual choice tests, the preferred drinkers were the bucket and the flowing water trough, but not the automatic drinking bowl (drinking time 25.2 ± 4.66, 11.5 ± 4.26, 2.4 ± 2.23 min for bucket, trough and bowl respectively, mean ± SEM; paired t-tests, bowl versus other drinkers, all tests p < 0.02). A possible reason for the avoidance of the automatic bowl was the noise it made when filling. After mixing a group of three ponies with a group of three unfamiliar animals, the ponies did not express their individual drinker preferences anymore. The use of the previously preferred bucket decreased significantly and the use of the initially, non-preferred, bowl increased significantly. This was likely caused by the fact that ponies were either intentionally or accidentally obstructing drinkers in certain areas of the pen and unfamiliar ponies did not want to push past them. In the simulated market, the differences in latencies to drink between ponies in the home pen and market groups did not reach significance. No significant effect of group size (groups of three versus six ponies) on drinking behaviour was detected. The results have implications for situations where only automatic water bowls are provided, such as during pony sales at livestock markets. Preventing ponies from expressing their drinking choice, either by offering non-preferred drinkers or by mixing with unfamiliar animals, could adversely affect their welfare especially if this happens in conjunction with other stressful events such as transport and markets, and potentially weaning.