Use of cues by grazing animals to locate food patches: An example with sheep

The ability of sheep to form spatial associations between two cues and two food rewards and to distinguish between the two cues prior to grazing them was examined, At 12 locations within a long, thin field, individual sheep chose between a pair of patches (plastic bowls) containing either preferred (cereal pellets) or non-preferred (lucerne, Medicago sativa pellets) food. Each patch had either a turf of perennial ryegrass (Lolium perenne) or white clover (Trifolium repens) associated with it as a cue, The cues were placed in wire mesh cages immediately in front of each patch. The cage meant the cue could be seen and smelled, but not eaten. Each sheep received a series of trials in one of three cue-patch associations: (i) 'WC', preferred patch always behind white clover and non-preferred patch always behind ryegrass; (ii) 'RG', preferred patch always behind ryegrass and non-preferred patch always behind white clover; (iii) 'R', preferred and non-preferred patches associated with cues at random. Five sheep were tested with each association, The spatial location of the food patches and cues was randomised within trials and rerandomised between successive trials to preclude animals from using spatial memory to locate patches. In the first trial, the proportion of times that the preferred food patch was chosen first at each pair of patches was similar in the WC (0.42), RG (0.46) and R (0.48) associations. In the sixth trial, the proportion of times that the preferred food patch was chosen first at each pair of patches was greater in the WC (0.81) and RG (0.82) associations than in the R association (0.46), The WC and RG data indicate that sheep formed associations between cues and rewards and distinguished between the two cues by sight or smell. The ability to form flexible associations between cues and food would be an important mechanism, acting independently of memory of spatial locations, that would help sheep select diets that are better than the average in spatially and temporally complex food environments. The ability to discriminate between white clover and ryegrass prior to grazing them indicates that the mixed diets (partial preference) exhibited by sheep given a choice between these two species do not simply arise because of a need for the sheep to sample the two species to know what each species is.