Burbidge & Burbidge (1959), Hodge & Merchant (1966), and Bertola & Capaccioli (1977, BC77) provided the first data for NGC 128 and its group. The galaxies of this group are late Ellipticals and early Spirals, and the recession velocities are in the range 4200-4600 km s^-1. NGC 128 is connected by a bridge to NGC 127, and according to BC77 a set of filaments protrudes southwards to NGC 126. The galaxy NGC 125 is surrounded by a large ring (BC77 estimated a diameter of $\sim$ 70 kpc) which is asymmetrically placed with respect to the center of the galaxy.

NGC 128 is classified as BS-I in the list of de Souza & dos Anjos (1987). The peculiar peanut shape of the bulge is clearly visible even in the innermost isophotes: the bulge appears squared by four symmetric bumps, forming an X-structure with arms at $\sim45^\circ$ from the major axis of the galaxy and extended, in projection, for $\sim$ 20'' (5.4 kpc). The peculiar morphology is not due to extinction effects (see next sections).

As suggested by Pfenniger & Friedli (1991) the X-structure is likely an optical illusion. We clearly see this effect in the original frame by changing the cuts: a thin disk and a small, and approximately round, bulge are seen in the center at the higher counts, while an increase of the thickness of the disk is apparent at lower counts. Such "flaring'' seems to originate the X-structure. The impression one has is that of looking to a "papillon''. The stars seem pulled out of the disk, the maximum effect taking place at $\sim$ 11 arcsec, corresponding to $\sim$ 3 kpc.

The galaxy is seen approximately edge-on. We derived a $\log(a_{25}/b_{25})$ ratio of 0.76, a value which is in relative good agreement with that found by Guthrie (1992) in his sample of edge-on galaxies for the S0 class. The observed major axis diameter is $\sim 50$ kpc (at $\mu_B=$ 26.0 mag arcsec^-2), and the thickness along the minor axis $\sim 21$ kpc.

The disk appears bended toward West on both sides, either in the visual and in the NIR images, in particular toward the South-West direction, where the bending starts at r = 37'' (10 kpc) from the center. The peaks of the light distribution along cuts perpendicular to the disk major axis have a maximum shift of 7'' (1.9 kpc).

$\begin{figure} \resizebox {\hsize}{!}{\includegraphics{ds7507f1.ps}}\end{figure}$

Figure 1: B-band maps of the O-C residuals for NGC 128: Upper panel: the result of the unsharp masking technique; Middle panel: the X-structure and the disk of NGC 128 after the subtraction of a model of the bulge component, obtained by fitting the inner isophotes with ellipses; Bottom panel: the result of the subtraction of a model of the bulge and disk component (see text). North is on the left, West is up. The image is $\sim$ 100'' in the N-S direction

We tempted to highlight the peculiar morphology of NGC 128 in different ways using the images with the best seeing and higher S/N ratio (the B-band frames). First we subtracted a model of the bulge component of the galaxy (Fig. 1, middle panel) obtained by fitting with ellipses the inner isophotal contours of the galaxy. In the residual map we identify the X-structure (in white color) and the disk component, which appears thicker in the outer region. The X-structure contributes approximately only to $\sim$ 5% of the total luminosity of the galaxy. This means that it is a projection effect rather than a physical new component.

While the X-structure and the bending of the disk are clearly visible in the original CCD frames, the increasing thickness of the disk in the outer region is possibly the result of the bulge subtraction (we verified this fact by simulating an edge-on S0 galaxy and subtracting the bulge component). Using the unsharp masking technique (Fig. 1, upper panel) we were able to confirm the bending of the disk on both sides, but not the flaring of the outer disk. The X-structure is obviously not visible in such image.

Notice that in Fig. 1 (middle and lower panels) the disk is fainter along the North direction. Here there is a strong absorption in correspondence of the encounter of the disk of NGC 128 with the arm of NGC 127.

Since the X-structure is the result of the subtraction of a model built with elliptical isophotes on a galaxy with a pronounced boxy shape, we realized a new 2D model of the galaxy taking into account the boxiness of the bulge. The 2D surface brightness distribution of the entire galaxy (within $\mu_B < 25$ mag arcsec^-2, masking the center, the region of the disk affected by the interaction with NGC 127, and the distorted outer disk) is given by the formulae:

	$\begin{eqnarray} \mu_{\rm b} &\!\!=\!\!& \mu_{\rm e} + k \times [(R_{\rm b}/R_{\... ... &\!\!=\!\!& \mu_0 + 1.086 \times (R_{\rm d}/h) \ \ \ {\rm (disk)}\end{eqnarray}$	(1)
		(2)

where $R_{\rm b} = \{[(x_i)]^c + [y_i/(b/a)_{\rm b}]^c\}^{(1/c)}$ and $R_{\rm d} = \{[(x_i)]^2 + [y_i/(b/a)_{\rm d}]^2\}^{(1/2)}$ are respectively the distance of the pixel (i,j) from the galaxy center. The exponent c>2 used for the bulge component realizes the boxy isophotes. The light profile of the bulge follows a r^1/n law (Caon et al. 1993), while the disk has an exponential distribution.

The model has the following best fitting structural parameters: $\mu_0 = 20.8\pm0.1$ mag arcsec^-2, $h = (24.5\pm0.5)''$ , $\mu_{\rm e} = 21.0\pm0.1$ mag arcsec^-2, $R_{\rm e} = (9.1\pm0.5)''$ , $(b/a)_{\rm b} = 0.67\pm0.05$ , $(b/a)_{\rm d} = 0.17\pm0.05$ , $n = 4.7\pm0.5$ , and $c = 3.69\pm0.05$ .This is, of course, only an approximated model for such peculiar galaxy having a strongly disturbed morphology. The $\chi^2$ fit appears infact too bright in the center, even after the convolution with the PSF of the image (Fig. 2) and, in the South direction, the true light distribution along the major axis is progressively fainter for r>40'', while along the minor axes the model is slightly brighter than the galaxy. Given the large uncertainty, we performed a second fit by eye, giving more weight to the less disturbed Southern region of the galaxy. This provided a smaller value for the scalelength h of 17''.

Even with the use of a more complex model, the residuals present an X-structure (Fig. 1, lower panel), a disk obscured in the North side, and a peculiar distorsion in the South-West direction. The X-component now contributes only to $\sim$ 2% of the galaxy luminosity. This proves that the X-structure is an optical illusion.

$\begin{figure} \resizebox {\hsize}{!}{\includegraphics{ds7507f2.ps}}\end{figure}$

Figure 2: The 2D models of NGC 128 with superposed the major and minor axes B-band light profiles (thick solid lines). The thin solid line is our $\chi^2$ solution, while the dashed line shows the fit realized by eye. North and East directions are in the left part of the diagram

In Figs. 3 and 4 we compare the structural parameters of NGC 128 with those extracted from a volume limited sample of elliptical, S0, and spiral galaxies of the Virgo and Fornax clusters (Caon et al. 1990, 1994 (C²D), D'Onofrio 1991). In terms of luminosity the galaxy belong to the "bright'' family of early-type objects defined by Capaccioli et al. (1992). The major and minor axes, measured by the parameters a₂₅ and b₂₅, are in good agreement with the corresponding data of the C²D sample for a galaxy of that luminosity (Fig. 3). The effective surface brightness (and the effective radius) of the whole galaxy and of its bulge component is relatively high: note in fact the peculiar position in the $\mu_{\rm e} - \log (r_{\rm e})$ diagram (Fig. 4 lower panel, cf. Capaccioli et al. 1992) which would assign the object to the "ordinary'' family of early-type objects.

For what concern the disk component, we measured the central surface brightness and the scale length of the disk and plotted them in the $\mu^c_0 - \log(h)$ diagram (Fig. 4 upper panel) comparing NGC 128 with a sample of 35 spiral galaxies of various morphological types (D'Onofrio 1991). The scale length of the disk appears very large while the corrected central surface brightness is normal for a galaxy of that luminosity.

**Table 1:** The data for NGC 128
$\begin{tabular} {llll} \hline\hline\noalign{\smallskip} Parameter & & Value & So... ...n et al.\/\ (1994]{coll:etal}) \\ \hline\hline\noalign{\smallskip}\end{tabular}$

$\begin{figure} \resizebox {\hsize}{!}{\includegraphics{ds7507f3.ps}}\end{figure}$

Figure 3: Position of NGC 128 in the $M_B-\log(a_{25})$ and $M_B-\log(b_{25})$ diagrams (lower and upper panels respectively). In the same plot we included for comparison the E and S0 galaxies of the Virgo cluster from the C² D sample

$\begin{figure} \resizebox {\hsize}{!}{\includegraphics{ds7507f4.ps}}\end{figure}$

Figure 4: Upper plot: The central surface brightness (corrected for inclination) and the scale-length of the disk of NGC 128 compared with a sample of 35 spiral galaxies of the Virgo cluster. The solid dots give the two results obtained from the fit of the galaxy for the scalelength h. Bottom plot: Position of NGC 128 in the $\mu_{\rm e} - \log (r_{\rm e})$ diagram. The solid dots show the values obtained for the whole galaxy and for the bulge component only. The error bars, non plotted here, are of the order of the dot sizes

The basic data for NGC 128 either derived in this work and extracted from the literature have been listed in Table 1. In the table $r_{\rm e}$ and $r'_{\rm e}$ are the radii of the circles which enclose half the total luminosity of the galaxy. The first was calculated by excluding the contribution of the two nearby companion galaxies of NGC 128. The HI mass has been calculated by us following Jura (1986).

2 Morphology