H I absorption was detected in all three galaxies close to the frequency
expected from the optical redshifts. Morganti et al. (2000) have independently
detected H I absorption in PKS 1814 -637. No emission was detected (or was
expected because of the redshifts). For PKS 1353-341 and PKS
1814-637, the absorption was both deep and of large velocity width.
The peak optical depths were 0.12 and 0.21, respectively; the
rest-frame velocity dispersions of the main absorption line were 66
and 19 km s-1, respectively; and the FWZI velocity extents were
496 and 372 km s-1, respectively. For PKS 1934-638, the optical
depth was about two orders of magnitude lower, =0.0022, and the
velocity dispersion of the absorption line was low, =8 km
s-1. Although very weak, the H I absorption for PKS 1934-638
was confirmed in both the ATCA and LBA observations. The results are
summarized in table .
Fig. shows the H I absorption spectra of the three sources.
The observations suggest that large quantities of H I must be present in 2 of the 3 galaxies: PKS 1353-341 and PKS 1814-637. Assuming a spin temperature T 102 K, the average column densities in front of the continuum sources are and atoms cm-2, respectively. For 1934-638, the column density is N(100)= atoms cm-2. This could indicate little neutral gas belonging to PKS 1934-638 in our line-of-sight. However, much higher column densities may be present if either: (a) the gas is warm ( could be K before the thermal linewidth approaches or more if radiative excitation of the spin levels is important); or (b) if there is a very extensive neutral gaseous halo. For example, Véron-Cetty et al. (1995) found a massive (3.1 1010 M) H I disk around the nearby compact flat-spectrum radio galaxy PKS 1718-649, whilst the optical depths of the two absorption components in this galaxy were only 0.003 and 0.005, similar to PKS 1934-638.
The Compact Array observations confirmed that all sources were unresolved to a maximum baseline length of 6 km, implying angular extents less than 5 arcsec. Further, the LBA observations showed that PKS 1934-638 has a similar flux density on all baselines up to 322 km, implying that it is unresolved (<01) to the accuracy of our calibration. This is in agreement with previous observations showing it to be a 42 milliarcsec double radio source (Tzioumis et al. 1989).