Imaging results

Fig.  shows the B and Ks images of the galaxy. The most conspicuous features in both images are the bar, the bulge and the spiral arms. The bar is better traced in Ks, since the distortions produced by the dust lanes are minimized. The circumnuclear ring and dust lanes are more clearly traced in B, where they appear as straight features from the bar edges towards the center, curving approximately within the bulge radius. All the ring features reported by Buta & Crocker (1993) are visible: (pseudo) outer, (pseudo) inner and circumnuclear rings, with sizes in agreement with those given by them, and also with optical and/or near-infrared images previously published (MM93, Friedli et al. 1996). We note the isophote twisting in the central parts, from the bar to the circumnuclear ring.

To parametrize the results of the photometric analysis, we have obtained the isophotal profiles by using the task ellipse within stsdas.analysis.isophote in IRAF, which uses the ellipse fitting method. All the stars in the frame were previously masked. The results are given in Fig. , where we plot isophotal magnitudes, ellipticity and position angle (PA) as a function of the square root of the radial distance for B, I, J and Ks bands. They agree with previous results by Friedli et al. (1996), who observed NGC 6951 with a similar configuration, although our data extends 30 arcsec further out. Since the area covered by our optical images is smaller than for MM93, and considering that the faintest isophotes we reach in NIR are still contaminated by the spiral arms, we have decided to use the disk major axis position angle and inclination as determined by MM93, i.e., 14#14 and i=42. These values are compatible with the outermost ellipses as traced by the sharp-divided images, and are in good agreement with the kinematically derived position angle and inclination, as we will see below.

To quantify the bar extension we have computed the amplitude and phase of the Fourier transform applied to the deprojected image, as explained in Márquez & Moles (1996). The mode m=2 has a constant phase out to 45, that we adopt as the bar extension.

The isophotal profiles indicate the presence of the circumnuclear ring and bar components, that appear as an excess around 5 for the former and from 20 to 50 for the second. These excesses are clearer once the photometric decomposition is made. We have fitted an exponential disk for the outermost radii, subtracted it from the profile, fitted an r^1/4 for the resulting bulge, subtracted it from the profile, and iterated until convergence is achieved. The results from the decomposition should be taken with caution since the bar contamination and the dust obscuration are very important. The results for the infrared image are presented in Márquez et al. (1999).

B-I and J-Ks full size colour images are shown in Fig. . In B-I the reddest features trace the dust lanes, most outstanding in the western side of the bar, while the bluest colours trace the regions of active star formation, that occur in the spiral arms outside a 30 arcsec radius. Within the part of the galaxy dominated by the bar structure there is very little ongoing star formation, except in the circumnuclear ring that is delineated by a number of blue knots. A comparison with the H1#1 image (fig. 5 in MM93; fig. 10 in González Delgado et al. 1997) shows that, for most regions, the H1#1 flux correlates with the colour B-I, however the bluest large region at 46 arcsec along position angle 20 does not have a correspondingly bright H1#1 emission, indicating that it may be a star forming region in the post-nebular phase. A detailed analysis of the stellar populations in NGC 6951 will be presented in a future paper. In the two colour maps, bulge and bar colours are very similar.

The highest contrast in the J-Ks image occurs around the western side of the circumnuclear ring, and along the western dust lane close to the ring. To look at these in more detail, Fig.  shows B-K and I-K colour maps, where K is the Calar Alto image, that has better sampling and seeing. Along the dust lanes, the colour contrast is higher in B-K, that results from a larger dust absorption in B than in I. For the reddest regions inside the circumnuclear ring, the situation is the opposite, with redder colours in I-K, in particular, close to the two ends of the ring major axis (NW-W and SE-S), and the SW zone where the ring connects with the bar. This higher contrast in I-K is similar to J-Ks in Fig. . The smaller contrast of the dust in B-K can be explained by an excess of B produced by the ongoing star formation in the ring. Thus the colour maps can be qualitatively understood in terms of the combined effects of dust obscuration and stellar population.

The HST WFPC2 V image has been published and analysed by Barth et al. (1995), that study the distribution of super star clusters in the circumnuclear ring. The HST NICMOS images have not been published yet. In Fig.  we show the J-H colour image. The red features contouring the circumnuclear ring are visible in great detail, mainly delineated as two short spiral arms that trace the ends of the oval. Closer to the nucleus, it consists of a number of small knots and structures that delineate a multi-spiral like structure.