Young migratory birds enter the world with two representations of the migratory direction, one coded with respect to the magnetic field, the other with respect to celestial rotation. The preferred magnetic direction of migratory orientation is malleable early in life: it may be calibrated by celestial rotation, observed either in daytime or at night. 2. Previous experiments showed that early experience with skylight polarization was necessary for calilbration to occur in daytime. In this study, we performed a direct manipulation of patterns of polarized skylight at dawn and dusk. 3. Hand-raised Savannah sparrows (Passerculus sandwichensis) were allowed to observe the clear sky for 1 h prior to local sunrise and for one h following local sunset. They never saw the Sun nor stars. The birds observed the sky through bands of polarizing material (HNP'B) aligned with the e-vector axis in one of three orientations with respect of the azimuth of sunrise and sunset: group 1) 90~ group 2) 45 ~ CW; group 3) 45 ~ CCW. 4. Tested indoors in covered cages in both shifted and unshifted magnetic fields, the autumn migratory orientation of the three groups differed significantly. Group 1 oriented magnetic N-S, group 2 oriented magnetic NW-SE, and group 3 oriented magnetic NNE-SSW. These observed orientation directions are very close to those predicted by the manipulations of polarized skylight. 5. These results indicated that a fairly simplified, static polarized light pattern viewed a limited number of times only in dawn and dusk snapshots is sufficient to produce calibration of the preferred magnetic migratory orientation direction. K. P. Able (1~)-M. A.