Nature 551, 619–622 (2017); doi:https://doi.org/10.1038/nature24646

In this Letter, the following paragraph was inadvertently omitted from the Methods section: ‘Gini coefficients for Tikal (Classic Maya) and Kahun (Middle Kingdom Egypt) were calculated from data compiled from site maps by the project ‘Service Access in Premodern Cities’ (http://cities.wikispaces.asu.edu/). A Gini coefficient for Mayapan (Postclassic Maya) was also calculated, using house size data provided by M. Masson and T. Hare. To our knowledge, none of these Gini coefficients has been published previously. A Gini coefficient for Caracol (Classic Maya) is taken from ref. 46. All of these measures are calculated from the areas of structures.’ This paragraph should have appeared in the ‘Statistical analyses and graphics’ section of the Methods, before the paragraph starting ‘Central Mexican Gini data are reported ...’. The original Letter has been corrected online.