EFFECTS OF THE SCALE AND SPATIAL-DISTRIBUTION OF THE FOOD RESOURCE AND ANIMAL STATE ON DIET SELECTION - AN EXAMPLE WITH SHEEP

1. The effects of the scale and spatial distribution of the food resou rce and animal state on diet selection of sheep were examined in two e xperiments. 2. In Experiment 1, sheep in indoor pens were allowed to f eed for 2 min from a patch (bowl) that contained a homogeneous mixture of 150 g of cereal (preferred) and 150 g of straw (non-preferred) she ep pellets. To test how the spatial scale (size) of individual food it ems affected the capacity for diet selection, both types of pellets we re cut to either 20, 15, 10 or 5 mm long (all 5 mm in diameter). To te st the effects of animal state on diet selection, the sheep were teste d after fasting for 0 (not fasted), 18 or 24 h. 3. In all treatments, sheep selected a diet that differed from the 50% on offer. The proport ion of cereal pellets in the diet was higher when pellet length was 20 mm than shorter pellet lengths, and when animals were not fasted than when animals were fasted for 18 or 24 h. 4. In Experiment 2, 100 equi distant patches (bowls) of food were laid out in a 50 m x 50 m vegetat ion-free field. Fifty were 'good' patches containing 105 g cereal pell ets and 45 g straw pellets, and 50 were 'bad' patches containing 45 g cereal pellets and 105 g straw pellets. To test whether the spatial di stribution of the food resource affected diet selection, patches were grouped together to create aggregations containing 1, 2, 5 or 10 patch es and these aggregations were randomly and independently assigned to positions in the field. Groups of three sheep were allowed to feed fro m the patches for 15 min, either after no fast or a 24-h fast. 5. Alth ough, in all cases, the proportion of cereal pellets in the total diet was higher than that expected if the sheep had foraged at random, the proportion was higher at the largest scale of aggregation (aggregatio n size of 10 patches) than smaller scales of aggregation and when shee p were not fasted. 6. The generally greater than expected proportion o f cereal pellets in the total diet arose because sheep ate from (i) a higher proportion of good patches, and (ii) a higher proportion of the cereal pellets within bad patches than expected from random. The incr eased selectivity seen at the largest aggregation size and when sheep were not fasted arose because these sheep ate from a higher proportion of the good patches. Within-patch selection did not vary between trea tments. 7. Overall, the results demonstrate sheep are able to discrimi nate between food items at a fine spatial scale, but that the scale of aggregation of alternate foods also affects diet selection. Diet sele ction may be further modified by changes in animal state brought about by fasting.