The character of land development affects the trip-making behavior of residents. The case is made for the careful demarcation of survey sampling districts from which to collect household travel data for forecasting travel demand. Household travel survey data from the Baltimore, Maryland, metropolitan area indicate that trip generation rates derived from areas with similar intensities of land development are more consistent than those derived according to political jurisdictions. Additional data from the Portland, Oregon, metropolitan area support this hypothesis. In light of current and previous findings, trip rates for metropolitan areas clearly should be derived from and applied to areas of homogeneous character in recognition of the effects of density and mixture of land uses on trip making. This assertion is illustrated with empirical data from the Baltimore metropolitan area. Conceptually, area type is more appealing than demarcations by political jurisdictions. Results of visual plots and numerical analyses show distinctive differences between the density classes but not as much between jurisdictions. A comparison of these results indicates that area type is the more superior form of stratification. Defining area type by density can reasonably capture future changes in land-use and activity patterns. The refinement suggested for the demarcation of sampling districts in the conduct of travel surveys is widely applicable to metropolitan areas around the country.