[1] Tract boundaries and population data for the 1950-1960 and 1960-1970 maps come from the National Historical Geographical Information System. In each map tract boundaries for the earlier year are used; population for the later year was assigned to the tract in which the centerpoint of the later year's tract fell. This procedure should have worked perfectly where tracts were split. It probably caused some minor errors in those relatively few cases where tract boundaries were shifted in more complicated ways. Nearly all of these instances occurred in the fast-growing suburbs.

[2] Only part of the area shown was "tracted" in 1950. The white spaces in the first map show these zones.

[3] Tract boundaries and population data for the 1970-1980, 1980-1990, and 1990-2000 maps were taken from the GeoLytics/Urban Land Institute Neighborhood Change Database, in which data are "normalized" to 2000 boundaries. Because the algorithm used often requires that data from thinly populated areas in the earlier years be distributed among several tracts, it is likely that there are some errors in the edge of the urban region. Population change maps for 1980-1990 and 1990-2000 that use boundaries from the earlier year (which may cause fewer problems) can be found elsewhere on this Website.