The linear response function technique is used to analyze two 1300-km tracks of SEASAT altimeter data and corresponding bathymetry in the Musician Seamounts region north of Hawaii. Bathymetry and geoid height are highly correlated in the 50- to 300-km wavelength range. A predictive filter is developed which can operate on SEASAT altimetry in poorly surveyed oceanic regions to indicate the presence of major bathymetric anomalies. Modeling of the bathymetry-geoid correlation in the Musician region is attempted using the elastic plate model. The flexural rigidity D of the plate is not well constrained by our data but appears to lie in the range 5¿1021 N m-5¿1022 N m at the time of loading. Since the Musician Seamounts and the crust on which they lie are both Late Cretaceous in age, this value represents the effective flexural rigidity of very young lithosphere that was 'frozen in' at the time of volcanism. The modeling indicates that the general form of a predictive filter will strongly depend on various geologic parameters, especially the effective flexural rigidity. Hence, some a priori geological constraints are necessary to estimate successfully the bathymetry from the altimeter data. Alternately, if high-quality bathymetry is available, a crude estimate of the age of loading (i.e., volcanism) can be made from the altimeter data.