Studies of sea level rise or ice sheet deformation from altimeter missions require careful calibration if systematic errors are not to be aliased into the result. For satellite missions that have distinct phases or when the repeat period is too long for repeat pass studies, altimetric crossover differences provide the mechanism for examination of residual systematic effects. In this study, altimetry, augmented with sea-surface variability time series as provided by the global network of tide gauges, is used to examine the stability of the ERS 1 range bias from April 1992 to August 1995. This 3-year period covered the first multidisciplinary phase, the geodetic phase, and the first four cycles of the second multidisciplinary phase. A correlation analysis using TOPEX/Poseidon and ERS 1 was first employed to identify a subset of island tide gauges where the time series is representative of that observed by the altimeter in the open ocean. Tide gauge augmented crossovers and all available repeat pass data were analyzed in the vicinity of these gauges. In particular, crossovers linking the ERS 1 multidisciplinary and geodetic phases permitted the range bias to be monitored continuously. Dual-crossover residuals between ERS 1 and TOPEX/Poseidon were utilized to gain insight into the characteristics of the temporal variation in the long-wavelength, geographically correlated error for the 3-year period. Failure to incorporate this geographically correlated error over the 3-year period led to a spurious drift of some 2--3 mm/yr. However, its inclusion reduces the drift to the extent that the observed rate is not statistically different from zero. ¿ 1999 American Geophysical Union