Cattle generalise visual cues from the pen to the field to select initial feeding patches

Free-grazing ruminants forage in environments containing multiple levels of complexity; the forage selection process operates at the landscape scale, when selecting feeding sites, and at the plant part level when selecting actual bites. Pen trials have shown that livestock associate visual cues with feeding sites, however, no field study has shown that animals generalise from training with visual cues in pens in order to choose feeding sites in the field. Our study tested nine beef heifers' ability to generalise a learned visual cue association to select feeding sites in a rangeland setting offering a novel forage, Lehmann lovegrass (LL). Initially, animals were trained to associate high quality feed with a visual cue during pen trials. We then tested animal response to the cue before and after they gained 14-day grazing experience with LL. Two identical field experiments (i.e., novel, before animals had experienced foraging LL versus familiar, after the animals had 14-day grazing experience with LL) were conducted over 3-day periods. Each experiment consisted of 27, 10-min trials. Animals were tested in plots containing high quality (HQ) and low quality (LQ) LL patches. For each trial, one of three randomly selected scenarios was presented: (1) the visual cue was placed in the HQ patch, (2) the visual cue was placed in the LQ patch, or (3) no visual cue was placed in either patch. Dependent variables were first patch-type chosen, bite rate in each patch, and number of observations of grazing in each patch. Cue presence influenced initial patch choice, bite rate, and grazing tallies within patch type. Heifers took 212 more HQ bites than LQ bites when the cue was placed in the HQ patch (P < 0.04), but took only 45 more HQ bites than LQ bites when the cue was placed in the LQ patch (P < 0.02). Heifers took 135 more bites from the HQ patch than the LQ patch when no cue was present (P < 0.02). Heifers clearly preferred HQ patches over LQ patches regardless of cue presence or absence, but grazed more in HQ and LQ patches when the cue was placed in those patches. The number of grazing tallies was directly related to bite rate within a patch. Animals grazed more in HQ than LQ patches when no cue was present. Visual cue placement altered this pattern; animals increased grazing in cued patches regardless of quality. Grazing experience did not influence observed grazing behaviour or the influence of the visual cue.