Zipf's law implies that a log-log plot of the rank-size distribution will be a straight line with a slope of negative one. A linear regression shows that the best fit to Zipf's law is found when close to 150 metropolitan areas are included.

The data for city sizes was obtained using the CityData function in Mathematica (September 2007), and the data for the metropolitan areas was obtained from the United States Census, 2000.