Abstract
Aim It has been widely assumed that large seeds generally require large animals to ingest and disperse them. However, this relationship has only been quantified in single animal groups (e.g. birds) and in a few communities. Our goal was to provide the first broad‐scale study of the relationship between animal body mass and ingested seed size. Location Global. Methods We compiled a dataset of 13,135 unique animal × seed interactions, animal body masses and seed sizes in these interactions, across all vertebrate groups (fish, amphibians, reptiles, birds and mammals). Results Contrary to expectations, ingested seed size was negatively related to animal body mass. This negative relationship was largely driven by large ungulates ingesting small and dry seeds, and analyses excluding either ungulates or seeds with non‐fleshy fruit types showed a positive relationship between animal body mass and ingested seed size. Large animals ingested both seeds with a larger maximum size (the 95th quantile had a positive slope) and a smaller minimum size (the 5th quantile had a negative slope). Larger animals ingest larger seeds from fleshy fruits but smaller seeds from non‐fleshy fruits. A significant positive relationship was found between animal size and the number of seed species ingested. Main conclusions Our data show that one of the assumptions that has underpinned the study of animal–seed interactions does not hold true across the full range of animal taxa and fruit types. These findings shed new light on theories about which types of plant species might be at risk if large animals go extinct, and cast doubt on the generality of a few theories (e.g. optimal diet theory, fruit‐size hypothesis) about the relationship between frugivores and seeds.