Astron. Astrophys. 326, 907-914 (1997)

5. Results

In the analysis of the PRs neighborhood, only the first three parameters are independent. They are: the number of objects, the cumulative diameter and the objects concentration. If a difference in the cumulative distribution of one of them exists, it will be present and amplified by the other three , , parameters.

The analysis of the parameters shows that there are no marked differences in the neighborhood of polar rings with respect to that of normal galaxies. This is shown in Table 3, where none of the significance levels for the parameters is above 85%. This is still valid in the sample of PR galaxies whose distance is known, labeled "Volume test" in Table 3. In this search of close companions, we excluded six galaxies which are too large to be recognized as a single object by the analysis software (used by APM or by FOCAS). This fact may bias the sample if they would represent extended or younger objects. However, as visible from Table 1, where these galaxies are identified by a , their linear size in Kpc is not different from that of the other PR galaxies in the sample. If objects like NGC 660 may represents young, unstable structures, on the other side IC 3370 or NGC 3384 have a very smooth appearance and may be older and dynamically relaxed.

In the analysis of the bright galaxies that may have encountered the PR before 1 Gyr, it is interesting to note the high frequency with which at least one galaxy of similar magnitude is found in the surrounding region. However, in comparison with the field of NGs, this fact is not statistically significant at a level of 91%. It is obvious that one must be careful in applying a statistical test to an analysis involving a single object. When a galaxy with a very similar red-shift lies near the polar ring, it is hard to think that a gravitational link between them is not present. The possibility that the present bright companions of PRs interacted with them must be analyzed for each single case.

In conclusion, the environment of PRs does not appear statistically different from that of normal galaxies. The number of fields surveyed in this paper is not large, but the result seems to be different from that reached for galaxies where interaction produces an observable nuclear activity, such as active galaxies or quasars (Dahari 1984, Heckman et al.  1985, Hintzen et al.  1991, Rafanelli et al.  1995), whose environments appear possibly richer than that of the normal galaxies.

Our data suggest that, if the event generating the PR was a mass transfer from a companion galaxy or a satellite ingestion, it should have happened in a remote epoch for the most part of galaxies and left almost no traces in the present. This idea seems supported by some arguments from this work and from the literature. First, the close environment around the PRs studied does not appear perturbed at the present epoch (Table 3). Second, some polar rings show a quantity of gas too high to derive from the ingestion of a single dwarf, late-type galaxy (Richter et al.  1994, Galletta et al.  1997). These massive rings are stabilized by the mechanism of self-gravitation (Sparke 1986). Their formation may have occurred in the early phases of the galaxy's life, when the number of late-type galaxies and their gas content were higher and the amount of accreted gas in a single encounter could have been large. Third, there is at least one galaxy of comparable size near almost all PRs (Table 4). With few exceptions these galaxies are not at present interacting with the PR, but they may have been gas donors in the past for the building of the ring. Finally, some models indicate ring formation times larger than 1 Gyr (Rix & Katz 1991 ) and/or a persistence of the ring until the end of the simulation (2.2 Gyr for Quinn 1991 and 7.2 Gyr for Rix & Katz 1991 ). A further support to this hypothesis may be given by the recent result of Reschetnikov (1997) on the detection of PRs in the Hubble Deep Field. He found that the number of PRs present in a 5 arcmin² field, 2 objects, is consistent with a PR space density increasing in the past.

The alternative explanations encounter some difficulties. The hypothesis that all the PR originates from the recent accretion of small satellites is not supported by the fact that many PRs are too massive to derive from the gas contained in a present-day dwarf galaxy. In addition, we may expect that an environment that is at present favoring the formation of a PR should be different from that of normal galaxies, which not seems confirmed by our results. The alternative that a PR forms by means of a slow infall of diffuse, primordial gas appears contradicted by the observations of emission lines in the rings. These are typical of gas regions observed in our Galaxy (CO, [N II], [SII], [OIII]) and show the presence of dust, which is a typical product of stellar evolution. In addition, it seems difficult for a slow infall to produce very inclined structures of low mass, because of the tidal torque of the host galaxy (Binney & May 1986).

In conclusion, the hypothesis that the majority of present PRs are 'fossil' structures born in the early Universe may be in agreement with the present data on the environments and with the presence of both massive and small PRs.

© European Southern Observatory (ESO) 1997

Online publication: April 8, 1998
helpdesk.link@springer.de