Distribution of linear polarization in Cassiopeia A at wavelengths of 9.8 and 11.1 cm

Two series of observations of the brightness distribution of the linearly polarized component of the radia- tion of Cassiopeia A are described. The first was made at a wavelength of 9.8 cm using a two-element interferometer at Stanford which provided a synthesized beamwidth of I `.6X2 7. The second set of ob- servations was made with the three-element interferometer of the NRAO at a wavelength of 11.1 cm and a beamwidth of 8'.' 1 X9'13 was obtained. Because the increments in the antenna spacing of the interferometer are greater than the critical interval for Cassiopeia A, part of the brightness distribution is missing from the synthesized map at 11.1 cm. This missing component is estimated to contain only 5% of the polarized flux density. The polarized radiation is concentrated in the main bright ring of the source and at 11.1-cm wavelength the maxima in the polarized brightness correspond to 5% of the unpolarized radiation. A comparison of the observations with published results at other wavelengths shows that Faraday depolarization reduces the polarized radiation to half the intrinsic value at a wavelength of approximately 7.5 cm. The mean values of the rotation of the position angle of the polarization at a number of points indicate a greater rate of rotation as a function of X2 at wavelengths shorter than 6 cm than at longer ories. These data are hardly detailed enough to rule out the possibility of linear rotation with 2 but appear to be best interpreted in terms of a nonlinear rotation within the source, together with a linear rotation of about -35 rad m which may be attributed to the interstellar medium. The internal rotation can be explained by a model of the source with a radial magnetic field and a thermal electron density of 2 . The Appendix discusses the effect on synthesis observations of the use of increments in the antenna spacing which are too large. Such undersampling results in the presence of grating lobes, the responses of which can only be separated from the main-beam responses for components of the source of dimensions less than the spacing of the grating lobes. Broader components of the source are lost, and to estimate the flux density of the unconfused main-beam responses it is necessary to correct the brightness scale by a factor derived from the beam pattern.