Analysis of the extent of segregation in the experimental systems was limited to a visual inspection of the top surface, due to the opacity of the granular bed. Videos were taken of the entire shaking process(N 3000 cycles) and counts of the number of the largest species visible on the bed surface were performed (Brazil nuts or green 11 mm
beads). A Brazil nut or large bead was counted to be on the surface if any portion could be discerned in the top view. Each Brazil nut is labeled with a‘B’ in the fi gures presented, to differentiate it from the similarly colored peanuts. The number of large particles on the top surface was used as a simple metric to quantify experimental
segregation. Discernable trends and patterns were captured using a similar metric in Schroter et al.[8]. We will also show later, in much more detail, that the trends in both simulations and experiments are in agreement