The impacts of disturbance were explored for red deer (<i>Cervus elaphus</i>) in the Scottish Highlands, by studying the distribution, behaviour and diet composition of deer in areas that were disturbed or less disturbed by human recreation, in different habitat types (grassland, heather moorland and woodland) and at different times of year.  The human-caused predation risk hypothesis states that there are similarities between the responses elicited by wild animals towards human recreational disturbance and predation.  This hypothesis was tested by comparing behaviour during the recreation season (spring and summer) and the hunting season (autumn and winter). Deer densities were lower in disturbed areas than less disturbed areas throughout the year, although there was no difference in relative habitat use as a result of disturbance. Habitat selection was primarily influenced by season, with grasslands having higher deer densities during the spring and summer and heather moorlands having higher densities during the winter. Behaviour was monitored using scan sampling of groups.  The percentage of animals that were vigilant was higher, resulting in a smaller number of deer feeding in disturbed than less disturbed areas during the recreational season.  Vigilance was higher in disturbed grassland and heather moorland than undisturbed woodland, while behaviour was similar in disturbed woodland and the less disturbed habitats.  In disturbed heather and woodland and in all less disturbed habitats, the majority of animals were standing while vigilant, whilst in disturbed grassland, lying was the main mode of vigilance.  Deer were more likely to be close together when vigilance levels were high.