3.1. The Photometry and Colors of Ring Galaxies
The first quantitative indication that ring galaxies were blue came from global U-B and B-V photoelectric color measurements made by Theys and Spiegel (1976) for 4 ring galaxies (VIIZw466, IZw45, IIZw28 and Arp 148). Colors were found to be in the range U - B = - 0.24 to -0.02 and B - V = 0.3 to 0.7. Lynds and Toomre (1976) presented colors for IIHz4, also finding the ring to be blue.
The most detailed early study of the colors of an individual ring, VIIZw466, was the work of Thompson and Theys (1978). The work was based on the calibration of KPNO 4m prime focus photographic plates through U, B and V filters. The main conclusion of the study was that all the knots in the ring have B - V colors < 0.4. The authors favored an explanation for the knots as being regions of young stars formed recently in a burst, although other constant star formation models were also considered as plausible. Figure 2b shows a peculiar loop or knot which extends inside this well defined ring. The knot appears much redder (B - V = 0.76) than the other knots, leading these authors to speculate that the knot might be the remnant of a displaced nucleus of the precursor galaxy (a possibility also conjectured by A. Toomre). Recent observations by Appleton and Marston (1995) do not support this idea. Indeed, the reddest region of the ring lies between the knot and the geometrical center of the ring. It is very likely that the knot is a star forming region.
Although the main result of the Thompson and Theys study was the extreme blue colors of individual knots, another interesting result emerged from their study, one which is now known to be a common property of ring galaxies, namely the hint of radial color gradients. In addition to the knots, the colors of three extended regions inside the ring were also measured and found to be redder in both U - B and B - V color than the knots. This result is confirmed by recent CCD observations by Appleton and Marston (1995).
Another indication of color gradients in rings came from the discovery of the high redshift ring (z = 0.24) from optical identifications of a deep radio survey (Majewski 1988). The ring's radio designation is 52W-036 (Windhorst et al. 1984) and it was recently observed as part of a driftscan CCD galaxy survey being carried out at Mt. Hopkins, and has the name KRN93-301 from that catalog (Kent, Ramella & Ninino 1994). Majewski observed the galaxy in UBVRI and the infrared K-band. This was the first published near-IR image of a ring galaxy. After applying a substantial K-correction the ring was found to be quite blue (U - B = - 0.79, B - V = 0.44) in line with other ring galaxies. Majewski noted that the morphology of the ring changed substantially from the blue, where it was crescent shaped, to the infrared, where it appears as a double source. The second source is probably the intruder galaxy. The change in morphology as a function of wavelength, appears more dramatic in 52W036 than that seen in the lower redshift rings (see below) and may indicate that it contains a substantial amount of dust.
Perhaps the most spectacular example of large radial color gradients is found in Zwicky's Cartwheel ring galaxy (Figure 1). Re-discovered by Lu (1971), the Cartwheel is perhaps the prototype ring galaxy. However, as we shall show later, it is far from representative of most ring galaxies in its global properties. In addition to the outer ring, there is a well defined inner ring (most easily seen in the near IR observations of Joy et al. (1988)). The Cartwheel was shown to contain extremely blue knots by the pioneering spectrophotometric study of Fosbury and Hawarden (1977). Strong optical and IR color gradients were found in the Cartwheel by Marcum, Appleton and Higdon (1992) and Higdon (1993). The change in color in the V - K color index was found to be over 2 magnitudes from the outer ring to the central nucleus (Figure 6). The colors are in approximate accord with a simple starburst model in which the stars are born in the ring with blue colors and evolve in the wake of the ring. These color changes are significantly larger than those found in normal spiral disks (typically no more than one magnitude in V - K, see de Jong 1995).
Figure 6. Figure 3 from Marcum, Appleton and Higdon (1992) showing the change in V - K color equally spaced annuli as a function of B - V color. The numbers refer to radius of the annulus, with 1 being the center and 8 and 9 being in the outer ring. Further results are also shown in Higdon (1993).
Recent observations of a sample of northern ring galaxies by Appleton and Marston (1995; see also Lysaght 1990) show that most of the larger ones observed exhibit radial color gradients similar to the Cartwheel. Figures 7a and 7b show two new examples, VIIZw466, the RN ring LT41 (Thompson 1977). A similar result is found for Arp 10. These figures clearly show the radial change of color, from blue in the outer ring to red in the center, pixel by pixel across the face of the galaxy. The radial color gradients provide perhaps the clearest and most beautiful evidence that stars are born suddenly in the rings and are left behind as the ring wave expands further into the disk. (Note: The possibility that dust absorption is responsible for the color gradients will be tested with IR observations with ISO.)
Figure 7. a) False color representation of
the B - R color change in the
VII Zw466 group. The colors represent changes in B - R
color from the
outer edge of the ring which are blue to the center which are
significantly redder. In the picture (as in Figure 7b), the reddest
regions in the galaxies are represented by yellow, the bluest by blue
(see color bars). For comparison, notice the red featureless colors of
the elliptical companion (Top). (See Color Plate IV at the back of
Observations also provide additional confirmation that the basic expanding ring picture is correct. For example, the simple kinematic picture of an off-center collision (e.g. Toomre 1978; Appleton and Struck-Marcell 1987b) predict a strong rarefaction behind the densest part of the ring. Such a rarefaction would have two observational consequences. Firstly, the surface density of the original target disk would be significantly reduced just inside the segment of the ring with the highest surface density. Indeed, the 2.2 µm IR emission is significantly reduced inside the bright southern segment of the Cartwheel (Marcum, Appleton and Higdon 1992). Secondly, the rarefaction should lead to a characteristic jump in velocity across the ring. In the leading edge of the ring, radial velocities are expected to be mainly directed outwards, whereas just inside the ring, the velocities should be infalling. In principal, the measurement of this velocity jump would be another diagnostic of the ring models. This second diagnostic is far more difficult to measure because it requires the presence of gas on both sides of the ring. Most of the ionized gas in ring galaxies is concentrated in the rings (Fosbury and Hawarden 1977; Taylor and Atherton 1984; Marston and Appleton 1995) and so is not useful for this kind of measurement. Cooler interstellar gas (either HI or molecular emission) holds the best promise for this kind of measurement. A hint of such a change in velocity across the ring is seen in Higdon's HI observations of the Cartwheel (Higdon 1993).
Twelve, mainly northern ring galaxies have been imaged by Appleton and Marston (1995) and Marston and Appleton (1995) through B, V, R broad-band and narrow-band H filters. In addition, J( 1.25 µm), H(1.65 µm) and K( 2.2 µm) near-IR images were obtained of the same sample. The main aim of the study is to investigate the optical-IR colors of the ring galaxies and to test models of star formation. The results provide the first systematic study of the global properties of ring galaxies.
The median B - V and V - K color of the ring galaxies in the sample are 0.52 and 2.31 magnitudes respectively, confirming the earlier results that the majority of the ring galaxies are blue and contain a substantial young population. The results show that the global B - V colors are not dependent on the linear size of the rings. On the other hand, there is a suggestion that when the optical to infrared baseline is used, a trend emerges. Larger ring galaxies appear to have larger V - K colors than small ring galaxies (see Figure 8). This result, which is consistent with the discovery of large radial color gradients, supports the view that larger ring systems are more evolved objects from a stellar evolutionary point of view. In the smaller ring galaxies the dominant emission comes from young stars in the ring. However, as we proceed to larger systems, the overall color of the galaxy becomes more and more dominated by the redder material inside the bright blue rings leading to the trend seen in Figure 8. This result suggests that, at least in some cases, any pre-existing stellar disk contributes only in a minor way to the overall luminous output of ring galaxies. This is probably only true in the case of the gas-rich systems, which dominate in samples of northern ring systems.
Figure 8. A plot of the optical to near-IR color (V - K) versus the linear diameter of the ring galaxy (from Appleton and Marston 1995).