Tim Newbold and Jorn Scharlemann recently published a paper using trait-based models of the response of bird species occurrence and abundance to land-use change to project the composition and diversity of whole bird assemblages in tropical forests. This paper shows that past land-use change has led to bird assemblages becoming - on average - less diverse, and also composed of a smaller abundance of certain guilds (nectarivores, frugivores and insectivores) and a larger abundance of others (especially generalist herbivores). The paper also generates future projections, showing that these past trends are likely to continue into the future under a business-as-usual scenario. While not directly a product of the PREDICTS project, the models underlying this paper used data from studies that have since been entered into the PREDICTS database. It also demonstrates how models like those being developed in the PREDICTS project can be used to infer changes in the structure and diversity of ecological assemblages.