Seed dispersal by endozoochory is important for the maintenance of plant populations and biodiversity. As a result, understanding the impact that frugivores’ activities have on seed dispersal is essential in order to better understand plant population dynamics. One factor that is known to affect an animal’s behaviour, yet has received little attention in this context, is animal cognition i.e. whether the information animals learn and remember affects where they access fruit and deposit seeds. Therefore, the aim of this thesis was to address how animal learning and memory affects the seed dispersal process, using two key approaches – experimental tests of frugivore cognition, and a model paramaterised to examine the consequences of different cognitive abilities on seed dispersal. Three questions were investigated: (1) The “where?” - whether the ability of frugivores to relocate previously visited food sources impacts upon their movements and, as a consequence, on plants’ seed shadows. The spatial learning and memory of red-footed tortoises was tested using an egocentric task. Tortoises were able to navigate efficiently in the environment, and remembered the spatial location of food for at least two months. A seed dispersal model designed to test whether frugivores with different spatial memory skills differently affect plants’ seed shadow, suggested that animals with long spatial memory relocate more efficiently food sources than animals’ with shorter memory. As a consequence, animals with longer memory survived longer, dispersed a bigger amount of seeds, and moved less at random around the environment, all of which lead to different spatial distribution of deposited seeds. (2) The “What?” - whether seed dispersers’ discriminatory skills and memory affects their choice of fruit. Tortoises’ ability to discriminate between quantity and quality of food was tested. They were able to successfully discriminate between the visual cues indicating different types of food and remembered the task for at least 18 months. A seed dispersal model designed to II investigate whether the memory of quality and quantity of food affects seed dispersal showed that the ability to discriminate between the features of fruits and, in particular, the memory of those, allows animals to base their foraging decision on previous learned experiences, significantly increasing the amount of seed dispersed from the preferred fruit. (3) The “When?” - whether the ability to anticipate events, such as food availability, and learn about plants fruiting cycles affects plants’ seed shadows. Tortoises’ anticipatory skills were tested on a 24 h cycle. They were able to anticipate food delivery, showing an increase in activity immediately prior the scheduled food delivery time. A seed dispersal model designed to test the impact of timing on dispersal showed that animals that are able to anticipate cycles equal to or longer than plants’ fruiting cycle readily relocate food, survive longer and disperse more seeds than those with memory that doesn’t last as long as a fruiting cycle. Finally, I present a model parameterised with red-footed tortoise cognitive data, with the aim to test one of the characteristics that makes tortoises unusual as seed dispersal vectors: the use of gaps in the forest. The results suggest that the active use of gaps enhances the probability of seed deposition in gaps and deforested areas, making tortoises a possible reforestation “tool”. I have demonstrated that the study of frugivores’ cognition can help to build more reliable predictions of seed dispersal by endozoochory: cognition is probably the most effective way to understand and predict an animal’s choices and movements around the environment. Future research should incorporate cognition in the study seed dispersal via endozoochory to have more reliable predictions of plant dynamics.