4 IR identifications and photometry

For the identification of the IR counterparts of radio sources we relied on the following informations: the most accurate radio positions available and the optical positions of stars visible in our IR frames. Concerning the radio positions, for the sources observed at the VLA we used the VLA coordinates of the radio core. In case the radio core was not detected we used the WENSS positions of the radio sources and we found it useful to overlay the VLA radio maps, when available, on our IR frames. For positioning the radio structure over our IR images, we needed astrometric information for our IR images. This was obtained using stellar objects in common between our images and R-band CCD images for which the astrometric work was already done (Rengelink & Snellen, private communication), or, when these were not available, using optical sky maps extracted from the digitized sky survey (DSS) plates available at the Space Telescope Science Institute.

The accuracy of our astrometric work depends on several factors: the uncertainties in the optical right ascension and declination positions; the errors in the radio positions; the accuracy in the radio/IR overlay procedure. The accuracy of the optical positions from the DSS plates is about 1 to 2 arcsec, depending on the distance of the object from the center of the Schmidt plate. The positional errors in the VLA radio maps are about 1 arcsec or better. The WENSS radio coordinates (used for some USS sources lacking VLA radio maps) have a 1 sigma error on the position better than 5 arcsec. For double sources that do not show an obvious radio core component, the extent of the radio source introduces an additional indetermination in the position of the IR counterpart. Finally, the accuracy of our radio/IR overlay procedure is about 0.5 arcsec. Taking into account all these uncertainties, we have accepted as good identifications those for which the distance between the radio and IR position is less than 1.5 to 5 arcsec, depending on the astrometric information available.

Tables 1, 2, 3 list the coordinates of the IR counterparts. In many cases, our IR detections provided us with an useful tool in order to confirm the identification of the fainter optical sources and to find the counterparts of the sources still lacking optical identifications. Figure 1 shows the K- and J-band images of the sources for which we have an IR counterpart, even if uncertain.

For all the IR counterparts, we measured the magnitudes using circular apertures with the minimum diameter which includes all the detected object flux. We tested this procedure by making several measurements with different aperture diameters. In a few cases we adjusted the photometric aperture to correspond to that used for the same object on the optical image by Snellen. The size of the apertures and the measured magnitudes, corrected for galactic foreground extinction, are listed in Tables 1, 2 and 3 for the three samples respectively. The photometric accuracy given in Tables 1, 2, 3 has been evaluated from the measured fluctuations of the sky background around the source for measurement apertures equivalent to that used for source itself. If no IR counterpart is detected we give 3$\sigma$ upper limit based on the background fluctuations. The J-K colours which we obtain for the USS sources and for the galaxies in the GPS sample are within the range of those observed for distant radio galaxies (e.g. Lilly & Longair 1984).

Figure 1: K- and J-band images (grey scale) of USS, GPS and FS radio sources. The arrows point to the IR counterparts. The crosses indicate the VLA positions of the radio cores or of the radio centroids; when available, the VLA radio contours are superimposed on the IR images. In the case of 0114+331 we show also contour levels of the IR flux, in order to enhance the two companions of the source at 4.5 arcsec to the south-east. In a few cases, bad pixels or columns are present, although they were corrected for the photometry

Figure 1: continued

Figure 1: continued

Figure 1: continued

Figure 1: continued

Figure 1: continued

Figure 1: continued

Figure 1: continued

Figure 1: continued

Figure 1: continued

Figure 1: continued

Figure 1: continued

Figure 1: continued