“…We can generate these rank size relations using various models: proportionate effect and preferential attachment of course, but also by subdividing a large hinterland into mutually exclusive subdivisions in a modular and regular manner, making various assumptions about population densities. This ties this kind of scaling back to both central place theory in the interurban context and to urban economics in the intraurban (Simon, 1955;Gabaix, 1999, Rozenfeld, Rybski, Andrade, Batty, Stanley, andMakse, 2008). In fact although rank-size scaling is highly stable through time, changes in the population of cities that make up such scaling can be highly volatile, and this remains a major puzzle in reconciling aggregate with disaggregate space-time correlations (Batty, 2010).…”