Abstract: Pulp of fleshy fruits may serve as a reward on seed dispersal by animals, but the effect of pulp nutritional value on seed removal speed has not been investigated. To quantify the effects of pulp nutritional value on seed removal speed, we conducted four experiments in the National Park of Hainan Tropical Rainforest, using depulped seeds, seeds with pulp, and artificial-pulp seeds with variable nutritional levels (carbohydrate-rich corn-based pulp and protein/fat-rich peanut-based pulp) across three tree species: Gnetum luofuense (Luofushan Joint-fir), Daemonorops jenkinsiana (Jenkins' rattan palm), and Dimocarpus longan (longan). We found intact D. longan seeds with pulp exhibited significantly higher seed removal speed (i.e. reversal value of seed survival time at the seed stations) than depulped seeds; and high pulp nutritional value (i.e. proportion of peanut mass to total artificial seed mass) significantly increased seed removal speed at seed stations, supporting the Optimal Foraging Hypothesis. However, contrasting to the hypothesis, we found pulp nutrition value showed a saturation effect on seed removal speed by small rodents, supporting our Pulp Reward Saturation Hypothesis, probably because animals would balance their nutrient demands (e.g. protein, fat and carbohydrate) by avoiding a single diet. Our study highlights the critical role of pulp in shaping seed-rodent interactions in tropical rainforest ecosystems.

Key words: Fleshy-fruit pulp, rodent, reward, mono-nutrient diet, saturation effect, seed removal, tropical rainforest