2. We collected data on voluntary food intake in mammals on forage-only diets and 22 related this to dietary neutral detergent fibre (NDF) content, assuming a nonlinear 23 correlation between these measurements. Generally, the paucity of corresponding data 24 is striking. 25 3. Elephants and pandas showed very high food intakes that appeared unrelated to 26 dietary fibre content. Only in small rodents, and possibly in rabbits, was an increase in 27 food intake on forages of higher NDF content evident. In particular, other large 28 2 herbivores, including horses, followed patterns of decreasing intake with increasing 29 forage NDF, also observed in domestic cattle or sheep. 30 4. For large herbivores, empirical data therefore do not -so far -support the notion that 31 intake is increased in response to declining diet quality. However, data are in accord 32 with the assumption that most large herbivores have an anticipatory strategy of 33 acquiring body reserves when high-quality forage is available, and reducing food 34 intake (and potentially metabolic losses) when only low-quality forage is available. 35 5. Intake studies in which the influence of digestive strategy on food intake capacity is 36 tested should be designed as long-term studies that outlast an anticipatory strategy and 37 force animals to ingest as much as possible. 38 6. We suggest that a colonic separation mechanism coupled with coprophagy, in order to 39 minimize metabolic faecal losses, is necessary below a body size threshold where an 40 anticipatory strategy (living off body reserves, migration) is not feasible. Future 41 studies aimed at investigating fine-scale differences, for example between equids and 42 bovids, should focus on non-domesticated species. 43