Our knowledge of how agriculture expands, and the types of land it replaces, is remarkably limited across the tropics. Most remote-sensing studies focus on the net gains and losses in forests and agricultural land rather than the land-use transition pathways (Gibbs et al 2010). Only a handful of studies identify land sources for new croplands or plantations, and then only for farming systems aggregated together (e.g., Koh and Wilcove 2008, Morton et al 2006, Gibbs et al 2010). Gutiérrez-Vélez et al (2011), however, have taken a leap forward by tracking the different expansion pathways for smallholder and industrial oil palm plantations