Several epidemiological studies have been based on wire codes (i.e., categories of electrical wiring configurations near residences) or on in-home spot measurements of magnetic flux density (MFD) as surrogates for short- and long-term exposure of children and adults to residential magnetic fields. We used wire code and MFD measurements that were made in 81 Colorado homes in 1985 and again in 1990 to assess their repeatability over periods of 0-24 h and 5 years. These homes, a subset of those lived in by subjects from the case-control study of Savitz et al. [Am J Epidemiol 128:21-38, 1988], were divided into four approximately equally sized groups that were differentiated by wire code and by case-control status. Eight homes were assigned wire codes in 1990 that differed from the 1985 coding of Savitz and colleagues [1988]; of these, seven were coded as high-current configuration (HCC) and one as low-current configuration (LCC) in 1985. Overall, 37 homes were coded as HCC in 1990 compared with 38 homes in 1985. Coding differences were due to differing distance measurements (four homes), differing "thick" vs. "thin" categorization of primary-distribution line-conductor sizes (two homes), differing "first-span" vs. "second-span" categorization of secondary wires (one home), and physical changes in proximate electrical wiring (one home). Coefficients of correlation between MFD spot measurements that were separated in time by 0-24 h range between 0.70 and 0.90. The coefficient between spot measurements made in 1985 and then again in 1990 is 0.70. These coefficients are similar for HCC and LCC homes and do not depend on whether residential appliances were turned on or off. The data show (at least for the portion of Colorado studied) that residential wire code and, more surprisingly, spot MFD measurements, are fairly reliable over 0-24-h and 5-year periods.