1999ApJ...518..676K

The Hubble Space Telescope (HST) archival UV imaging polarimetry data of NGC 1068 are reexamined. Through an extensive estimation of the observational errors, we discuss whether the distribution of the position angles (P.A.s) of polarization is simply centrosymmetric or not. Taking into account the effect of a bad focus at the time of the observation, we conclude that, within the accuracy of HST/FOC polarimetry, the P.A. distribution is completely centrosymmetric. This means that the UV polarization originates only from scattering of the radiation from a central pointlike source. However, our analysis shows that the most probable location of the nucleus is only ∼0".08 (∼6 pc) south from the brightest cloud called cloud B. The error circle of 99% confidence level extends to cloud B and to cloud A, which is about 0".2 south of cloud B. By this FOC observation, cloud B is only marginally rejected as the nucleus. Assuming that the UV flux is dominated by electron-scattered light, we have also derived a three-dimensional structure of the nuclear region. The inferred distribution suggests a linear structure that could be related to the radio jet.