Fig. 1 shows the HI (in green) and CO(1-0) (in black) integrated column density distributions, superimposed on an optical V band image obtained at the CFHT. The most striking result is the marked spatial separation between the atomic hydrogen gas and the molecular gas as traced by the CO emission. Almost no HI is detected towards the spiral whereas this galaxy contains high quantities of molecular gas. HI emission is predominantly seen along the northern tidal tail, and to a smaller extent along the southern filament. In what follows, we will discuss the gas distributions in more detail.
3.1. The HI gas
The observational results of our HI observations are presented in the following figures: in addtion to Fig. 1, Fig. 2 displays the individual profiles of the 21 cm line in the southern part of the system; Fig. 3 shows the individual channel maps of the 21 cm line, separated by intervals of 20 km s-1, superimposed on the optical V-Band image; and Fig. 6 presents the HI velocity map. Table 4 summarizes the HI integrated properties of the system: the angular and linear sizes of the HI clouds (Col. 2,3), the peak column density (Col. 4), the HI masses (Col. 5), derived using the formula by Giovanelli & Haynes (1988) , and the FWHM of the 21 cm line (Col. 6), integrated over the entire cloud extent.
Table 4. VLA results
3.1.1. The spiral galaxy
No HI is found coinciding with the spiral galaxy itself, except perhaps for a small cloud of mass , at the detection limit of the VLA. It is located to the NW of the disk next to a star forming region.
3.1.2. The northern tail (Cloud A105N)
As shown in Fig. 1, the bulk of the atomic hydrogen is distributed along the northern optical tidal tail. The peak of emission is situated on the Magellanic irregular galaxy (A105N), and more precisely at a distance of 110 kpc from the nucleus of NGC 3561A. It coincides with optical knots which were found to be HII regions (Paper I). This cloud will be referred hereafter as Cloud A105N. The HI distribution is highly asymmetric. Getting closer to the spiral, the HI column density gradually decreases, dropping below at a distance of 30 kpc from the nucleus of the spiral. The total HI mass of Cloud A105N, , is slightly higher than that determined from single dish Arecibo observations ( , Paper I). Fig. 3 reveals unexpected and complicated features in the HI kinematics. In A105N, several components at different velocities coexist along the line of sight. This explains, in particular, the unusually high HI velocity dispersion in Cloud A105N it has a mean value of 25 km s-1 and peaks at 70 km s-1 in the diffuse feature, west of the Magellanic Irregular , and makes the determination of a single mean velocity from the peak of the HI emission line (Fig. 6) rather ambiguous. Nevertheless, the velocity determined towards the HII regions in A105N matches that measured in the optical (see Table 7). We will look into the detailed kinematics of this arm in Sect 4.2.1.
3.1.3. The southern tail (Clouds N3561B and A105S)
The southern part of Arp 105 shows two distinct systems, one of which is seen in emission, Cloud A105S, and the other in absorption, Cloud N3561B. The former coincides with the blue compact object A105S, the latter with the elliptical galaxy NGC 3561B. Because of the limited spatial resolution of the VLA C+D configuration ( ; kpc), and of the small separation between A105S and NGC 3561B ( ; ), it is only because of their difference in velocity (170 km s-1 ) that it is possible to differentiate the two.
Cloud A105S is resolved by the VLA beam; it is elongated in the South-North direction, along the thin optical filament that reaches A105S, and located at a distance of 50 kpc from NGC 3561A. It can be traced in emission up to a distance of , north of the nucleus of NGC 3561B (see Fig. 2), which means that, taking into account the beam size, it extends up to at least 15" (8.5 kpc) south of NGC 3561B. The derived mass of Cloud A105S is , which is one tenth of the mass of Cloud A105N. The optical and radio velocities towards the nucleus of A105S are comparable; the difference between the two is 30 km s-1. However, although a steep velocity curve in that object had been determined from optical H observations (Paper I), the mean velocity of the HI cloud appears to be roughly constant over its total extent (2 beam sizes).
Cloud N3561B is situated in front of the elliptical galaxy NGC 3561B, which is known to host a radio continuum nucleus with a flux density of 42.4 mJy at 20 cm. This radio source, against which HI is seen in absorption, is unresolved at a linear resolution of 180 pc ( Batuski et al. (1992) ). The absorption line spectrum is presented in Fig. 4. The velocity of the peak of the HI absorption is redshifted by 250 km s-1 with respect to the velocity of the nucleus, determined from optical longslit spectroscopy (Paper I). Moreover the spectrum appears asymmetric; its blue wing is extending all the way to the optical velocity. A likely interpretation for this is that HI between us and the elliptical galaxy is falling towards it.
In the intermediate region, between the two clouds, the HI spectra show both emission and absorption features (see Fig. 2), at two separate velocities. In between these velocities, there is a "plateau" where no HI is detected. In Sect. 4.2.2, we will discuss the origin of the differences in the velocity distribution and try to determine whether Clouds A105S and N3561B really form a single structure or not.
3.1.4. Other HI detections
Apart from Arp 105, three other HI sources, in the velocity range 8400-9000 km s-1, were detected in the VLA field. Their positions, integrated fluxes and derived velocities, are indicated in Table 5. As we explain in Appendix A, we made additional observations of an object at a velocity around 6250 km s-1 which, based on its optical appearance, appeared to be interacting with NGC 3561A (Paper I). In the end, it proved to be unrelated to Arp 105.
Table 5. HI detections in the VLA field
3.2. 20 cm radio continuum results
Fig. 5 displays the VLA 20 cm continuum results for a field around Arp 105. This map was based on the line-free channels of the combined VLA C- and D-array dataset. The rms noise is about 0.08 mJy. Both the spiral as well as the elliptical are strong continuum sources. Their flux densities are and mJy, respectively. Two extended sources are clearly visible as well, around position ; , near the centre of the cluster Abell 1185, and another one south of it. The latter one is likely related to the elliptical galaxy NGC 3554 with which it coincides. The northern object, or cloud, is very extended and diffuse with no obvious optical counterpart. It coincides with the X-ray gas as mapped by Beers et al. (1991) . Such objects have been seen now in a few clusters of galaxies (see Feretti & Giovannini (1996) , for a review). A follow-up study is in progress.
3.3. The molecular gas
Based on two single dish CO pointings with the IRAM 30-m antenna (Paper I), we already could determine which regions of the system contained molecular gas: no CO was detected within a diameter region centred on the Magellanic Irregular, A105N. In contrast, a huge amount of molecular gas ( , estimated using an N (H2 ) to I (CO) ratio of , and the formula of Sanders et al. (1984) ) was found in a similar region centred on the spiral. Fig. 1 shows high resolution follow-up observations with the Plateau de Bure array in the 12 CO(1-0) line towards the spiral. Table 6 summarizes the properties of the molecular gas in NGC 3561A.
Table 6. IRAM Plateau de Bure results
It is clear from these interferometric observations that the molecular gas, as traced by the CO molecule, is concentrated in the central region of the spiral galaxy. All the emission is found inside a radius of (2.8 kpc) centred on the nucleus. The CO peak itself lies at less than (0.4 kpc) from the optical nucleus. Given our astrometric precision, which is also , there is no evidence for any offset of the CO distribution with respect to the spiral nucleus. Combes et al. (1988) have mapped the molecular gas content of the Virgo galaxy NGC 4438. They showed that, due to the interaction with a companion galaxy, of the gas had been displaced by ( 5 kpc, at the distance of the Virgo cluster). We do not observe such a clear effect in Arp 105. However given the large distance of this system and the sensitivity achieved by the PdB, we would not have detected off-centered clouds less massive than (10% of the total mass). Besides the CO peak emission coincides with the position given by Batuski et al. (1992) for the centre of the centimetric continuum emission to an accuracy of better than (see Table 1).
The total CO flux measured with the Plateau de Bure interferometer matches very well (5% error) that obtained inside the beam of the 30-m antenna, if one corrects the latter for its beam efficiency. This means that the PdB array has missed no flux.
The distribution of molecular gas in NGC 3561A is particularly smooth and symmetric, apart from a faint East-West extension. The optical image shows, at a larger scale, a similar structure. This faint extension ends to the West in an HII region, and to the East coincides with a diffuse EW feature.
The velocity map (Fig. 6b) of the spiral, derived from the CO(1-0) line observation, shows a barely resolved gradient of 180 km s-1 kpc-1 inside a nuclear radius of 1 (0.6 kpc), and a flat rotation curve outside. Fig. 7 presents the velocity curves of NGC 3561A along two position angles: along the South-North direction (PA =0, shown in the main figure), and along a direction closer to the major axis of the galaxy (PA = -35, shown in the insert in Fig. 7). It combines the HI and CO results from the present paper with the optical velocities derived from H longslit spectroscopy (Paper I). All velocities used in this figure are calculated based on the optical Doppler shift definition to allow a direct comparison between the various wavelength ranges. The optical and CO rotation curves of NGC 3561A agree in their overall shape but there are small differences. In particular, on the south-east side, the CO velocities are higher than the optical ones by 70 km s-1. This could be due to the fact that CO and H do not have the same optical depth, and therefore do not arise in the same region. In the direction of the elliptical, we did not detect any CO emission to a limit of 3 Jy km s-1. An unresolved continuum source of 30 mJy is found at the location of the 20 cm radiosource and against which HI is detected in absorption (Table 6).
© European Southern Observatory (ESO) 1997
Online publication: October 15, 1997