1996A&A...315L...5M

We have measured the 800µm polarization at three positions towards S140 and at the flux peak towards GL 2591. For both sources the implied magnetic field direction is neither parallel nor perpendicular to the outflow axis, implying an "in-between" situation. A review of published submillimetre polarimetric observations of outflow sources implies this scenario is common. Whilst beamsize effects are not thought to be critical, comparison of these observations with the results from a recent theoretical model (Bonifacio & Emerson 1995, ASP Conf. Ser., 97, 363) imply that the "in-between" situations may actually be due to the effect of viewing angle on the observed polarization position angle. If the local magnetic field is uniform and perpendicular to the outflow axis the observed position angle of polarization for dichroic emission of aligned grains is highly dependent on the viewing angle. The surprising agreement between mid-infrared and submillimetre emissive polarization position angles for W3-IRS5, NGC 7538-IRS1 and OMC1-IRc2, and their non-orthogonality to the mid-IR absorptive polarization position angles, has implications for the magnetic field configuration around these outflow sources. In particular, an hour-glass magnetic field geometry may be applicable.