There are two contradictory approaches to explaining the presence of secondary metabolites in ripe fruits. One holds that they evolved toward enhancing dispersal success (adaptive approach); the other claims that they evolved primarily to deter herbivores from eating leaves and seeds and that their presence in ripe fruits is a byproduct of that function (non-adaptive approach). We tested the validity of three hypotheses of the adaptive approach that explain the presence of secondary metabolites in ripe fruits. We explored the current function of a secondary metabolite, emodin, in Mediterranean buckthorn (Rhamnus alaternus, Rhamnaceae) fruits by relating intraspecific variation and seasonal patterns of concentration to fruit removal and seed damage and by conducting feeding trials with captive birds presented artificial fruits that varied in emodin concentration. The concentration of emodin in the pulp of 10–13 Rhamnus plants from the same population was determined by HPLC every month during two fruiting seasons. Fruit removal by birds and seed predation by invertebrates and microbes were determined for the same plants. Emodin concentration rose during the first stages of ripening, reaching a peak before the fruits were ripe, and then decreased to a minimum when the fruits were ripe. No significant correlation between emodin concentration and ripe fruit removal rate among trees was observed in the first year, whereas in the second year the correlation was positive and significant. Thus, the impact of emodin on fruit selection varied between years, suggesting that emodin concentration does not solely govern fruit selection. A significant negative correlation was found in the first year between emodin concentration and seed predation during the first fruiting month. The yellow-vented bulbul (Pycnonotus xanthopygos), a seed dispersing bird, distinguished between artificial foods that differed in emodin concentration (control, 0.001% and 0.002%), always preferring the lower concentration. In contrast, house sparrows, (Passer domesticus), a seed predator, did not detect such differences in emodin concentration but did distinguish between control foods and food with 0.005% and 0.001% emodin. We suggest that emodin has an ecological role, preventing seed predation by invertebrates and microbes without decreasing fruit removal by avian dispersers.