Low availability of functional seed trait data from the tropics could negatively affect global macroecological studies, predictive models and plant conservation.

BACKGROUND: Plant seeds have many traits that influence ecological functions, ex situ conservation, restoration success and their sustainable use. Several seed traits are known to vary significantly between tropical and temperate regions. Here we present three additional traits for which existing data indicate differences between geographical zones. We discuss evidence for geographical bias in availability of data for these traits, as well as the negative consequences of this bias. SCOPE: We reviewed the literature on seed desiccation sensitivity studies that compare predictive models to experimental data and show how a lack of data on populations and species from tropical regions could reduce the predictive power of global models. In addition, we compiled existing data on relative embryo size and post-dispersal embryo growth and found that relative embryo size was significantly larger, and embryo growth limited, in tropical species. The available data showed strong biases towards non-tropical species and certain families, indicating that these biases need to be corrected to perform truly global analyses. Furthermore, we argue that the low number of seed germination studies on tropical high-mountain species makes it difficult to compare across geographical regions and predict the effects of climate change in these highly specialized tropical ecosystems. In particular, we show that seed traits of geographically restricted páramo species have been studied less than those of more widely distributed species, with most publications unavailable in English or in the peer-reviewed literature. CONCLUSIONS: The low availability of functional seed trait data from populations and species in the tropics can have negative consequences for macroecological studies, predictive models and their application to plant conservation. We propose that global analyses of seed traits with evidence for geographical variation prioritize generation of new data from tropical regions as well as multi-lingual searches of both the grey- and peer-reviewed literature in order to fill geographical and taxonomic gaps.