Residential floor dust loading was measured on the smooth floor surface of 488 houses in Syracuse, New York, during the summers of 2003 and 2004. Using U.S. Environmental Protection Agency (EPA) wipe methods, pre-weighed Ghost Wipes, Lead Wipes, or Whatman Filters were employed to collect duplicate samples from (predominantly) kitchens. The collection efficiency of the various media was determined from multiple wipe tests and side-by-side comparisons. The results were normalized and aggregated at the census tract level to determine whether spatial patterns of dust loading could be observed. Loading was found to be log-normally distributed, with a geometric mean value of 0.311 g m(-2) (29 mg of dust per square foot of floor); 95% of the observations fell in the range of 0.042-2.330 g m(-2) (4-216 mg foot(-2)). The sampling for floor dust loading shows some bias for day of the week in which visits to the residential properties were made. After a first-order correction for this effect, results were aggregated by census tract and mapped in a geographic information system (GIS); strong spatial patterns can be identified in an inverse distance weighted mapping. The geographic patterns exhibit a strong correlation with socio-economic/demographic covariates extracted from the 2000 census summaries. Dust mass on the floors is positively correlated with renter-occupied properties and family size; it is negatively correlated with measures of household income.