In principle, there are three ways that such elongated polarization minima could be formed:
The first interpretation is unlikely because the decrease in polarized intensity P at the rails is often 30% to 50% of the polarized emission from the lobe on adjacent sightlines. Improbable geometries (slab-like features in the lobes with their long axes aligned with the line of sight) would therefore be needed to generate low-polarization regions on both sides of the jets. The second interpretation is highly unlikely because the fractional depth and separation of the rails changes little between 1.4 and 8.4 GHz, and because our four-frequency data reveal no unusual Faraday rotation features near the jet (Swain 1996).
The third interpretation, vector cancellation, is compatible with all of our data. The apparent magnetic field direction inferred in the lobes, after correction for Faraday rotation, is approximately perpendicular to the jet axis along most of the length of both jets. If sightlines through the outer layers of these jets, but not those through their centers, are dominated by parallel to the jet axes, then the net polarized intensity can have minima near the jet edges, as observed.
Figure 3: The correlation between rail depth and jet intensity measured at the rail minima at 8.4 GHz with FWHM resolution wherever the rail depth exceeds .
Figure 3 shows that the "rail depth'' (the difference between the local minimum in P and the value interpolated from the ambient polarized emission), correlates well with the total intensity I of the jet at the position of the minimum, everywhere the rail depth exceeds 3. (The jet total intensity was determined by fitting and removing the lobe background from each profile using a low-order polynomial). The only badly discrepant point comes from the south edge of knot J1, where a rail is detected but the intensity profiles differ significantly from the mean. We infer that:
The rail minima never go to zero, so the polarized jet emission never completely cancels that from the rest of the line of sight. In the few places where the polarized intensity between the rails is positive with respect to the ambient polarization, the apparent magnetic field near the jet axis must therefore be perpendicular that axis, to reinforce the lobe polarization.
We now present a simple model for the jet that is consistent with these results.