Roles of novelty, generalization, and postingestive feedback in the recognition of foods by lambs

Ruminants select foods higher in nutrients and lower in toxins than the ave rage of what is available whether fed in confinement or foraging on pasture s or wild lands. Our objective was to better understand how sheep learn to select the most nutritious foods when they concurrently ingest different fo ods. We hypothesized that novelty and generalization are two ways sheep dis criminate among foods. The first experiment determined whether lambs offere d two foods in a meal (one a novel-flavored food and the other a familiar f ood) acquired a preference for the novel-flavored food following an infusio n of energy (starch) into the rumen. Lambs did not increase preference for the novel-flavored food when the amount of starch infused (150 g) was held constant (P > 0.05). However, when the amount of starch infused was made di rectly proportional to the amount of novel-flavored food ingested, lambs qu ickly formed a preference for the novel-flavored food (P < 0.001). The seco nd experiment determined whether lambs generalized preferences from familia r to novel foods. Lambs fed coconut-flavored mile grain for 51 d subsequent ly preferred coconut-flavored straw to plain straw (P < 0.001). Conversely, lambs that were not fed coconut-flavored mile grain preferred plain to coc onut-flavored straw (P < 0.001). Lambs infused with a source of energy (sta rch, 100 g/[lamb.d]) consumed more coconut-flavored straw and less plain st raw than lambs that had no infusions or lambs infused with a toxin (LiCl, 1 00 mg/kg BW; P < 0.05). Lambs infused once with LiCl had the lowest prefere nce for coconut-flavored straw (P < 0.05). Lambs conditioned with starch ha d the highest intakes of coconut-flavored straw, even after starch infusion s were suspended (P < 0.1). When straw was replaced by an energy-dense food (wheat grain), all lambs equally preferred coconut-flavored grain to plain grain (P < 0.001). Collectively, our results suggest that lambs learned to prefer a novel-flavored food when the amount of energy reward was continge nt on the amount of novel-flavored food ingested, and that they generalized from familiar to unfamiliar foods based on common flavor cues. Once genera lization occurred, postingestive feedback from nutrients and toxins calibra ted preference according to the food's utility.