2 Observations and data analysis

2.1 Observations

The 53 observed galaxies suspected to have a "transition'' spectrum are listed in Table 1 with the various names under which they are known, and in Table 2 with the published line intensity ratios $\lambda 5007/{\rm H}\beta$, $\lambda 6300/{\rm H}\alpha$ and $\lambda 6583/{\rm H}\alpha$. Table 3 gives their optical positions measured on the Digitized Sky Survey (Véron-Cetty & Véron 1996).

Spectroscopic observations were carried out during several observing runs in May, June and July 1996 and January, March, October and November 1997 with the spectrograph CARELEC (Lemaître et al. 1989) attached to the Cassegrain focus of the Observatoire de Haute-Provence (OHP) 1.93 m telescope. The detector was a $512 \times 512$ pixels, $27 \times 27 \,\mu$m Tektronic CCD. We used a 600 lmm^-1 grating resulting in a dispersion of 66 Åmm^-1; the spectral range was $\lambda\lambda\,6700-7600$ Å in the red (with a Schott GG 435 filter) and $\lambda\lambda\,4860-5760$ Å in the blue. In each case, the galaxy nucleus was centered on the slit.

Usually five columns of the CCD ($\sim$ 5$^{\prime\prime}$) were extracted. The slit width was 2$.\!\!^{\prime\prime}$1, corresponding to a projected slit width on the detector of $52 \,\mu$m or 1.9 pixel. The slit position angle was not always the same for the blue and red spectra; as the aperture used is rectangular (2$.\!\!^{\prime\prime}$1 $\times$ 5$^{\prime\prime}$), this may introduce some inconsistencies when the line emitting regions are extended. The resolution, as measured on the night sky emission lines, was $\sim \! 3.4$ Å FWHM. The spectra were flux calibrated using the standard stars given in Table 4, taken from Oke (1974); Stone (1977); Oke & Gunn (1983) and Massey et al. (1988). The journal of observations is given in Table 5.

2.2 Line profile fitting

Morgan (1958, 1959) has introduced a classification of galaxies based on their nuclear region stellar population. Classes "a'' and "af'' are dominated by early-type stars. The main absorption features are the Balmer lines, which are usually filled up by emission as these objects invariably contain a H II region. Classes "g'', "gk'' and "k'' are dominated by a bulge of old population II stars. Intermediate classes "f'' and "fg'' have, in addition to a population of young stars, a faint bulge of old stars. The old star population have similar spectra in all classes (Bica 1988).

 

Table 1:  Cross-reference names for the galaxies studied in this paper

 

Table 2:  Published line intensity ratios for the 53 emission-line galaxies studied in this paper. References: (1) Aguero et al. (1995), (2) Anton (1993), (3) Augarde et al. (1994), (4) Boller et al. (1994), (5) Cohen & Osterbrock (1981), (6) de Grijp et al. (1992), (7) González Delgado & Perez (1996), (8) Duc et al. (1997), (9) Fruscione & Griffiths (1991), (10) Goodrich & Osterbrock (1983), (11) Hill et al. (1988), (12) Ho et al. (1997a), (13) Keel et al. (1985), (14) Kim et al. (1995), (15) Klaas & Elsasser (1991), (16) Kollatschny et al. (1983), (17) Koratkar et al. (1995), (18) Koski (1978), (19) Martel & Osterbrock (1994), (20) Netzer et al. (1987), (21) Osterbrock & Pogge (1987), (22) Osterbrock & Martel (1993), (23) Phillips et al. (1983), (24) Rafanelli et al. (1990), (25) Salzer et al. (1995), (26) Shuder & Osterbrock (1981), (27) Ulvestad & Wilson (1983), (28) Veilleux & Osterbrock (1987), (29) Véron et al. (1997), (30) Vogel et al. (1993)

 

Table 3:   B1950 optical positions of the observed objects measured on the Digitized Sky Survey. The rms error is 0$.\!\!^{\prime\prime}$6 in each coordinate; "*'' indicates objects with larger errors due to their location near one edge of the Schmidt plate (Véron-Cetty & Véron 1996). References for finding charts: (1) Andreasian & Alloin (1994), (2) Arp (1966), (3) Bowen et al. (1994), (4) Carballo et al. (1992), (5) de Grijp et al. (1987), (6) González Delgado & Pérez (1996), (7) Duc et al. (1997), (8) Kazarian (1979), (9) Keel (1996), (10) Markarian & Lipovetski (1971), (11) Markarian & Lipovetski (1973), (12) Markarian & Lipovetski (1974), (13) Markarian et al. (1979a), (14) Markarian et al. (1979b), (15) Markarian & Stepanian (1983), (16) Mazzarella & Boroson (1993), (17) Olsen (1970), (18) Pesch & Sanduleak (1983), (19) Rubin et al. (1975), (20) Sandage & Bedke (1994), (21) Takase & Miyauchi-Isobe (1986), (22) Takase & Miyauchi-Isobe (1990), (23) Vogel et al. (1993), (24) Wyndham (1966)

AGN activity is exceptional in classes "a'' and "af'' but frequent in all other classes (Véron & Véron-Cetty 1986); in consequence, the nuclear region of most AGNs contains a population of old stars with many strong absorption lines which can make the line fitting analysis rather difficult, especially for the blue spectra. Therefore, when the absorption blend Mg Ib $\lambda$5175 was prominent, we subtracted a suitable fraction of the spectrum of the elliptical galaxy NGC 5982, used as a template, to remove the old stellar population contribution. Another elliptical, NGC 4365, was used as a template for the red spectra, while NGC 821 was used for both the red and blue spectra obtained in October and November 1997. All templates were observed with the same spectral resolution as the emission-line galaxies.

Whenever necessary, we have added a $\rm H\alpha$ or $\rm H\beta$ absorption component; as, usually, the $\rm H\alpha$ absorption line is completely filled up by the emission lines, we assumed its intensity to be 1.8 times the intensity of the nearby absorption Ca I$\lambda$6495 line (Véron-Cetty & Véron 1986b). Whenever a template and/or absorption component was used in a fit, this is indicated in Table 6 which contains the line fitting analysis results for the 53 observed galaxies.

The line fitting analysis of the spectra was done in terms of Gaussian components as described in Véron et al. (1980, 1981b,c). The three emission lines, $\rm H\alpha$ and [N II]$\lambda\lambda$6548, 6583 (or $\rm H\beta$ and [O III]$\lambda\lambda$4959, 5007) were fitted by one or several sets of three Gaussian components; whenever necessary, two components were added to fit the [O I]$\lambda\lambda$6300, 6363 lines. The width and redshift of each component in a set were supposed to be the same. The intensity ratios of the [N II]$\lambda\lambda$6548, 6583, [O III]$\lambda\lambda$4959, 5007 and [O I]$\lambda\lambda$6300, 6363 lines were taken to be equal to 3.00, 2.96 and 3.11, respectively (Osterbrock 1974). Whenever a fit resulted in a small negative intensity for a $\rm H\beta$ component, we set the corresponding $\lambda 5007/{\rm H}\beta$ ratio to 10, the mean value for Seyfert 2s. All line widths given in this paper have been corrected for the instrumental broadening. The spectra and best fits are plotted in Fig. 1, the parameters describing the individual components required by the analysis being given in Table 6.

2.3 Notes on individual objects

Mark 938. This galaxy is apparently undergoing a merger as evidenced by the presence of tidal tails (Mulchaey et al. 1996; Mazzarella & Boroson 1993). The nature of its emission-line spectrum has been rather controversial. Afanasjev et al. (1980) classified it as a Seyfert 2, Osterbrock & Dahari (1983) claimed that it is not a Seyfert; Dahari (1985), Véron-Cetty & Véron (1986a) and Veilleux et al. (1995) called it again a Seyfert 2, while Mulchaey et al. (1996), observing a weak emission of [O III]$\lambda$5007 and a strong $\rm H\alpha$ over the entire galaxy, suggested that there is no Seyfert activity in this object, in agreement with Mazzarella & Boroson (1993) who called it a H II region. The line ratios published by Veilleux et al. (1995) and Vaceli et al. (1997) indicate a "transition'' spectrum, the [O III] lines being weak for a Seyfert 2 galaxy ($\lambda 5007/{\rm H}\beta$ = 4). The high resolution spectroscopic observations of Busko & Steiner (1990), showing complex emission line profiles with great differences in width and shape between $\rm H\alpha$ and [N II]$\lambda$6583 (the measured line widths are $264 \pm 7$ and $384 \pm 12$ $\rm km\, s^{-1}$ for $\rm H\alpha$ and [N II]$\lambda$6583, respectively), suggest a "composite'' spectrum. To fit our spectra, two components are needed: one is a H II region with narrow lines ($\sim$ 255 $\rm km\, s^{-1}$ FWHM); the other is a Seyfert 2 with much broader lines ($\sim$ 760 $\rm km\, s^{-1}$ FWHM). For this component, we find a very high and unlikely $\lambda 6583/{\rm H}\alpha$ ratio ($\sim$ 6.8); however, there is a very strong and broad $\rm H\beta$ absorption line. It is probable that the broad $\rm H\alpha$ emission component intensity is greatly reduced by the presence of a $\rm H\alpha$ absorption line which has not been accounted for.

Mark 957. This galaxy has been identified with the radiosource 5C 3.100 (Antonucci 1985) and a ROSAT X-ray source (Boller et al. 1998). Dahari & de Robertis (1988) called it a Seyfert 2. However, Koski (1978) and Halpern & Oke (1987) have observed strong Fe II emission lines in this object; furthermore, the continuum is very flat, extending far into the blue (Koski 1978), accounting for the classification of this object as a Narrow Line Seyfert 1 galaxy. This classification is supported by Boller et al. (1996) who have found a steep soft X-ray component (photon index $\Gamma = 2.9 \pm 0.2$) with a variable flux (by a factor 1.9 over 18900 s). The $\rm H\beta$ line is very narrow (FWHM < 685 $\rm km\, s^{-1}$) (Goodrich 1989); narrow $\rm H\alpha$ and [N II] lines are observed as far as 10$^{\prime\prime}$ from the nucleus (with $\lambda 6583/{\rm H}\alpha$ $\sim \!0.4$) (Halpern & Oke 1987), suggesting the presence of an extended H II region. In the nucleus, the high ionization lines ([O III] and [Ne III]$\lambda$3869) are found to be blueshifted by $\sim$ 280 $\rm km\, s^{-1}$ with respect to the low ionization lines. Although having a relatively low signal-to-noise ratio, our spectra are quite interesting. In the blue, there is a very narrow $\rm H\beta$ emission line ($FWHM \sim 200$ $\rm km\, s^{-1}$) associated with very weak ($\lambda 5007/{\rm H}\beta$ $\sim$ 0.15) and relatively broad ($FWHM \sim 710$ $\rm km\, s^{-1}$) [O III] lines; the associated broad $\rm H\beta$ component is weak ($\lambda 5007/{\rm H}\beta \sim 9$) and accounts for only 8% of the total $\rm H\beta$ flux. The [O III] lines are blueshifted by $\sim 360$ $\rm km\, s^{-1}$ with respect to $\rm H\beta$. The red spectrum is also reasonably fitted with two sets of components; one is narrow with weak [N II] lines, while the second is broader with relatively strong [N II] lines. This is in satisfactory agreement with Halpern & Oke's results, and suggests that the nuclear spectrum is dominated by a strong H II region superimposed onto a relatively weak Seyfert 2 nucleus.

IRAS 01346 - 0924 was identified by de Grijp et al. (1987) with a galaxy they called MCG -02.05.022, which seems to be erroneous. It was classified a Seyfert 2 by de Grijp et al. (1992) on the basis of its emission- line ratios. We discussed this object in Paper I, giving it the wrong name (MCG -02.05.022); we suggested, on the basis of a blue spectrum, that it was "composite''. The best blue spectrum fit is obtained with three sets of three Gaussians, two being typical of a H II region and the third of a weak Seyfert 2 nebulosity.

 

Table 4:  Spectrograph settings and standard stars

 

Table 5:  Journal of observations. A: 66 Åmm^-1, red; B: 66 Åmm^-1, blue. A "*'' indicates objects published in Paper I

 

Table 6:  Fitting profile analysis results. Column 1 gives the name of the object, Col. 2 the adopted redshift, Cols. 4 and 9 the velocities for each set of components measured on the blue and red spectra, respectively, and de-redshifted using the redshift given in Col. 2; Cols. 5 and 10 the corresponding FWHM, Cols. 6, 11 and 12 the intensity ratios $\lambda 5007/{\rm H}\beta$, $\lambda 6583/{\rm H}\alpha$ and $\lambda 6300/{\rm H}\alpha$ respectively, and Cols. 7 and 13 the fraction of the $\rm H\beta$ emission flux (respectively $\rm H\alpha$) in each component with respect to the total flux of the line in each object. A "T'' in Col. 3 (or 8) indicates that the blue (or red) spectrum has been corrected for starlight using a suitable fraction of a template (in the blue, we have used the elliptical galaxy NGC 5982 and in the red, the elliptical galaxy NGC 4365; NGC 821 was used as a template for the objects observed in October and November 1997); in the same columns, an "A'' indicates that a $\rm H\beta$ (or $\rm H\alpha$) absorption component was added to the model. In Col. 14 we give the velocity difference between the blue and red systems, and in Col. 15 the spectroscopic classification of each component in the model; Gaussian profiles were used throughout, except when indicated by "lz'' (Lorentzian profile). Values in parenthesis have been fixed

 

Table 6: continued

 

Table 6: continued

 

Table 6: continued

A weak broad ($FWHM \sim 2\,640$ $\rm km\, s^{-1}$) $\rm H\alpha$ component may also be present. The Seyfert 2 cloud is so weak that it is not detected on our red spectrum.

Mark 1066 is an early-type spiral galaxy (Afanasjev et al. 1981; Mazzarella & Boroson 1993). It was classified as a Seyfert 2 by Afanasjev et al. (1980) and as a Seyfert 1.9 by Afanasjev et al. (1981) on the basis of weak broad components in the Balmer lines; the existence of these broad components has not been confirmed by Goodrich & Osterbrock (1983) who concluded, on the basis of the line ratios ($\lambda 5007/{\rm H}\beta$ = 4.35, $\lambda 6583/{\rm H}\alpha$ = 0.88, $\lambda 6300/{\rm H}\alpha$ = 0.08), that this object is a Seyfert 2. However, Osterbrock & Dahari (1983), on the basis of the same data, called it a "marginal'' Seyfert 2, the $\lambda 5007/{\rm H}\beta$ ratio being relatively weak. The spectra published by Wilson & Nath (1990) and Veilleux (1991a) show that the emission lines have a broad blue wing extending up to velocities of 1000 $\rm km\, s^{-1}$ with respect to the line peaks, the [O III] lines being significantly broader (403 $\rm km\, s^{-1}$ FWHM) than the Balmer lines (280 $\rm km\, s^{-1}$) (Veilleux 1991b,c). De Robertis & Osterbrock (1986) noted a good correlation between the width of the line at half-peak intensity and the critical density, suggesting that the narrow line region (NLR) is density stratified; however, the density stratification mostly affects the high velocity gas producing the wings of the line profiles (Veilleux 1991c). Haniff et al. (1988) have published an [O III] image suggestive of a double structure with a separation of 0$.\!\!^{\prime\prime}$8, and the continuum nucleus in between. There are two emission peaks in the core of the low-ionization lines; these peaks are separated by about 0$.\!\!^{\prime\prime}$5, the velocity difference between them being $\Delta V = 125 \pm 20$ $\rm km\, s^{-1}$ (Veilleux 1991c). High-resolution (0$.\!\!^{\prime\prime}$1) HST images (Bower et al. 1995) reveal that the $\rm H\alpha$ and [N II] emission comes from a 3$^{\prime\prime}$-long region centered on the nucleus, while the [O III]-emitting gas is concentrated in a bright "jet-like'' structure extending 1$.\!\!^{\prime\prime}$4 NW of the nucleus. Long-slit spectroscopy suggests the existence of two kinematically distinct regions: the first, of low-excitation, lies in the plane of the galaxy and is normally rotating, while the second, of high-excitation, would be inclined with respect to the disk. Bower et al. (1995) suggested that the high-ionization cloud is a Seyfert 2 (with $\lambda 5007/{\rm H}\beta$ $\sim \!10-15$) and the low-ionization cloud is a Liner ($\lambda 5007/{\rm H}\beta$ $\sim\! 2-3$); it seems, however, that the [O I] lines are weak and that this region could be a H II region instead. Three sets of lines are needed to fit our spectra; one set originates, most probably, from a H II region, the two others having line ratios typical of Seyfert 2 clouds. The velocity difference between the two Seyfert components is $\Delta V = 146$ $\rm km\, s^{-1}$ in the blue and 105 $\rm km\, s^{-1}$ in the red. The complexity of the line emission regions in this object, revealed by the HST observations, makes the measured line ratios for each individual component rather inaccurate. It seems likely that the density-stratified cloud emitting relatively broad lines is compact and coincides with the nucleus. This is, therefore, a "composite-spectrum object''.

IRAS 03355 + 0104 has been identified by de Grijp et al. (1987) with a galaxy shown to be a Seyfert 2 by de Grijp et al. (1992) who have measured $\lambda 6583/{\rm H}\alpha$ = 0.58, a normal value for such an object; however, Vogel et al. (1993) have found much weaker [N II] lines, with $\lambda 6583/{\rm H}\alpha$ = 0.18. Our red spectrum gives $\lambda 6583/$H$\alpha = 0.49$, in agreement with de Grijp et al. (1992), and $\lambda 6300/{\rm H}\alpha$ = 0.12, so this object is a Seyfert 2 galaxy. In addition, a weak broad $\rm H\alpha$ component seems to be present, in which case it would be a Seyfert 1.9 galaxy.

IRAS 04210 + 0400 has been identified with a compact blue galaxy with a faint blue, spiral companion (Moorwood et al. 1986). It is associated with a double lobed radio source, 20-30 kpc in size (Beichman et al. 1985; Hill et al. 1988). The galaxy has an apparent spiral structure (Beichman et al.); however, these features are dominated by emission lines, and the galaxy is probably an elliptical (Hill et al. 1988; Steffen et al. 1996).

 

Table 7:  HST observing log of IRAS 04210+0400

We have searched the Hubble Space Telescope archives and found images obtained with the Wide Field Planetary Camera 2, on January 31, 1995 through medium and broad band filters isolating several emission lines and a line-free continuum. We retrieved and analysed these unpublished images, the HST observing log being given in Table 7. The galaxy was imaged on the Planetary Camera, a $800 \times 800$ pixels CCD with a readout noise of $\sim 5$ e^-pixel^-1. The pixels size is $15 \times 15\, \mu$m, which corresponds to 0$.\!\!^{\prime\prime}$0455 on the sky; the field is 36$.\!\!^{\prime\prime}$4 $\times$ 36$.\!\!^{\prime\prime}$4 (Trauger et al. 1994; Holzman et al. 1995).

Both the $\rm H\alpha$+[N II] and the [O III] images (after subtraction of the continuum) show a very complex structure with a bright unresolved nucleus, a relatively bright elongated central region extending over $\sim$ 2$.\!\!^{\prime\prime}$4, made of several distinct clouds, and a thin spiral feature with a total extent of about 15$^{\prime\prime}$ (Fig. 2).

The Balmer decrement observed over a 2$^{\prime\prime}$$\times$ 2$^{\prime\prime}$ area centered on the nucleus is relatively large ($\rm H\alpha$/$\rm H\beta$ = 5.5) (Hill et al. 1988), implying a large extinction ($A_{V} \sim 1.4$ mag). We cannot exclude the possibility that the extinction varies over the emission nebulosity; therefore, the ratio ($\rm H\alpha$+ [N II])$\lambda$5007, which is approximately equal to $\rm H\alpha$$\lambda$5007 as [N II]$\lambda$6583 $\sim$ 1/3 $\rm H\alpha$, cannot be taken as an estimate for the excitation parameter $\lambda 5007/{\rm H}\beta$. Our entrance aperture (2$.\!\!^{\prime\prime}$1 $\times$ 5$.\!\!^{\prime\prime}$0), with the slit oriented in the E-W direction, basically includes the central point source and the bright central nebulosity.

Published nuclear line ratios (Hill et al.) led to the classification of this object as a Seyfert 2 (Beichman et al. 1985; Holloway et al. 1996) although $\lambda 6583/{\rm H}\alpha$ = 0.35, a low value for this class. The core region shows asymmetric spatial structure with several separate components in velocity and space; there are systematic shifts between peak positions for the different lines (Holloway et al. 1996). Our spectra basically confirm the line ratios obtained in the nuclear region. This is a Seyfert 2 galaxy with abnormally weak [N II] emission lines.

IRAS 04507 + 0358 has been identified with an elliptical galaxy (de Grijp et al. 1987) shown to be a Seyfert 2 by de Grijp et al. (1992) who, however, have measured $\lambda 6583/{\rm H}\alpha$ = 0.28, a very low value for an object of such a class. Our red spectrum gives $\lambda 6583/{\rm H}\alpha$ = 0.47 with, perhaps, a weak broad Lorentzian $\rm H\alpha$ component. Therefore, this is not a "composite-spectrum object'', but rather a Seyfert 2 galaxy or perhaps a Seyfert 1.9, if the broad component is confirmed.

VII Zw 73 has been classified as a Seyfert 2 galaxy by de Grijp et al. (1992) on the basis of its emission line ratios ($\lambda 5007/{\rm H}\beta$ = 3.96, $\lambda 6583/{\rm H}\alpha$ = 0.56); however, as in the case of Mark 1066, the [O III]$\lambda$5007 line is rather weak for a Seyfert 2. Our blue spectrum shows $\rm H\beta$ to be clearly narrower than the [O III] lines, an indication of the probable "composite'' nature of this galaxy. To obtain a good fit to our blue and red spectra, three components are needed. However, two of these components turned out to have similar velocities and widths, making the determination of the line ratios rather uncertain. We therefore re-observed this galaxy with a higher resolution (33 Åmm^-1) in the red on October 29, 1997 and in the blue on October 30. Again three components were needed to fit the spectra, but this time they were clearly identified as corresponding to two Seyfert 2 nebulosities and a H II region.

Mark 622. The line ratios published by Shuder & Osterbrock (1981) show that it is a Seyfert 2 galaxy, although the [O I]$\lambda$6300 relative flux is too low for such a class ($\lambda 6300/{\rm H}\alpha$ = 0.06). Furthermore, these authors found that the [O III] lines are much broader ($FWHM \sim 1\,050$ $\rm km\, s^{-1}$) than the $\rm H\alpha$, [N II] and [O II] lines ($FWHM \sim 350$ $\rm km\, s^{-1}$) suggesting the presence of two emission-line regions of different ionization. Wilson & Nath (1990) have shown that, in the nucleus, the [N II] lines are broader than $\rm H\alpha$ ($FWHM = 340 \pm 20$ and $240 \pm 20$ $\rm km\, s^{-1}$ respectively); moreover, the [O III] image of this object is only slightly resolved, while $\sim$ 60% of the $\rm H\alpha$ emission comes from an extended component (Mulchaey et al. 1996). Our spectra confirm these results; in fact, two components are needed in order to obtain a good fit: one is representative of a H II region, the other is typical of a Seyfert 2 cloud.

3C 198.0. The line ratios in this elliptical radio galaxy are those of a H II region, as discussed in Paper I. Our measurements are in agreement with the published values: $\lambda$6300/$\rm H\alpha$ = 0.05 and $\lambda 6583/{\rm H}\alpha$ = 0.28. This is therefore a really puzzling object.

KUG 0825 + 248. The published emission-line relative intensities (Augarde et al. 1994) are typical of a H II region except for the [O I] lines, which are too strong ($\lambda 6300/$ $\rm H\alpha$ = 0.13). On our red spectrum, we measured $\lambda$6300$/\linebreak$$\rm H\alpha$ = 0.02. Our values for $\lambda 5007/{\rm H}\beta$ and $\lambda 6583/{\rm H}\alpha$ are in agreement with the published values. This object is, therefore, a typical H II region.

IRAS 09111 - 1007 has been identified with the western component of a galaxy pair (Murphy et al. 1996). The eastern component, separated by 40$^{\prime\prime}$, is called IRAS 09111-1007E, although it is probably unrelated to the IR source; it has a "transition'' spectrum with $\lambda 5007/$H$\beta = 3.94$, $\lambda 6583/{\rm H}\alpha$ = 0.74 and $\lambda 6300/{\rm H}\alpha$ = 0.07 (Duc et al. 1997). Our observations suggest that the spectrum of this object is indeed "composite'', the [O III] lines being broader than $\rm H\beta$. The blue spectrum CCD image clearly shows the $\rm H\beta$ line to be double and spatially resolved. In fact, to obtain a good fit to the $\rm H\alpha$+[N II] and $\rm H\beta$+[O III] lines, three components are needed: one with line ratios typical of a Seyfert 2 cloud, and the two others typical of H II regions.

Zw 238.066. On the basis of the published line intensity ratios, Veilleux et al. (1995) have called this object a Liner; however, the [O I] lines are weak ($\lambda 6300/{\rm H}\alpha$ = 0.07). Our observations suggest that it has a "composite'' nature. The red spectrum image clearly shows the $\rm H\alpha$ and [N II] lines to be extended, with a low relative intensity of the [N II] lines. The blue spectrum shows that the [O III] lines have a broad blue wing not seen in $\rm H\beta$. We have therefore fitted both the red and blue spectra with three sets of lines. The blue spectrum has a rather poor signal-to-noise ratio which probably explains why the parameters of the blue and red fits (especially the line widths) are not in good agreement; however, the differences of profile of the different lines leave no doubt about the "composite'' nature of this spectrum. Indeed, the fits show that the emission-line spectrum is dominated by two H II clouds separated by $\sim$ 220 $\rm km\, s^{-1}$; but there is, in addition, a weak Seyfert 2 nebulosity with relatively broad lines ($\sim 1\,030$ $\rm km\, s^{-1}$ FWHM).

UGC 5101. This is an ultra-luminous infrared galaxy ($L\rm_{IR}$ $\geq$ 10¹² $L\rm_{\hbox{$\odot$}}$); it is a peculiar galaxy with a large ring and a 15 kpc tidal tail extending to the west, which suggests a merger, although no companions are known to this galaxy. It has a single bright nucleus (Sanders et al. 1988). Optical spectra have been published by Sanders et al. who classified it as a Seyfert 1.5 galaxy on the basis of a relatively broad $\rm H\alpha$ line, by Veilleux et al. (1995) and Wu et al. (1998a,b) who called it a Liner, and by Liu & Kennicutt (1995). However, the published line ratios ($\lambda 5007/{\rm H}\beta$ = 2.9-4.7, $\lambda 6583/{\rm H}\alpha$ = 1.17-1.35, $\lambda 6300/{\rm H}\alpha$ = 0.06-0.09) rather indicate a "transition'' spectrum. Our red slit spectrum (PA = 270$^\circ$) shows spatially extended narrow emission lines with a steep velocity gradient across the nucleus in addition to broader, spatially unresolved lines in the nucleus itself. We have extracted five columns centered on the nucleus and analysed separately the five spectra. In each case, we have found a narrow component ($FWHM \sim\! 100-150$ $\rm km\, s^{-1}$) with $\lambda 6583/{\rm H}\alpha$ $\sim$ 0.50 and $\lambda 6300/{\rm H}\alpha$ $\sim$ 0.05, the velocity decreasing from $\sim$ 250 (West) to -120 $\rm km\, s^{-1}$ (East). On three spectra, we detected relatively broad ($FWHM \sim 500$ $\rm km\, s^{-1}$) lines, with $\lambda 6583/{\rm H}\alpha$ > 1 and $\lambda 6300/{\rm H}\alpha$ < 0.40. In addition, on the two central spectra, there is a weak, broad ($FWHM \sim 1\,200$ $\rm km\, s^{-1}$) $\rm H\alpha$ component. It was not possible to perform such a detailed analysis on the blue spectrum, which has a much lower signal-to-noise ratio due to a large Balmer decrement ($\rm H\alpha$/$\rm H\beta$ = 8.3; Sanders et al. 1988). Nevertheless, we can draw some conclusions: the medium width $\rm H\alpha$ component flux, coadded on the three central columns, represents 24% of the total $\rm H\alpha$ flux on the same three columns (excluding the broad $\rm H\alpha$ component); assuming that the Balmer decrement is the same for the narrow and medium width components, and that the narrow component is a H II region (this component having $\lambda 6583/{\rm H}\alpha$ $\sim$ 0.50, must have $\lambda 5007/{\rm H}\beta$ < 2), we conclude that the intermediate width lines set must have $\lambda 5007/{\rm H}\beta$ > 3.5 and is, therefore, a Seyfert 2 cloud. So, UGC 5101 has a "composite'' spectrum with a rather strong starburst component and a Seyfert 1.8 or 1.9 nucleus (we are not able to decide between 1.8 or 1.9 as, because of the rather poor signal-to-noise ratio around $\rm H\beta$, it is not possible to put a significant upper limit to the broad $\rm H\beta$ component flux).

NGC 2989 is a Sc galaxy (Sandage & Bedke 1994). Published data (Table 2) indicate a "transition'' spectrum; Phillips et al. (1983) called it a intermediate object, but Véron-Cetty & Véron (1984), on the basis of the same line ratios, classified it as a H II region. Our measured flux ratios ($\lambda 5007/{\rm H}\beta$ = 1.70, $\lambda 6300/{\rm H}\alpha$ = 0.04 and $\lambda 6583/$ $\rm H\alpha = 0.53$) are in good agreement with Phillips et al. and show that this is indeed a pure H II region with, perhaps a marginally large $\lambda 6583/{\rm H}\alpha$ ratio.

CG 49. This galaxy has been shown to have a Seyfert 2 spectrum by Salzer et al. (1995); however, they have measured a low relative intensity for the [N II] lines ($\lambda 6583/{\rm H}\alpha$ = 0.30) as the redshifted wavelength of [N II]$\lambda\,6583$ is 6873 Å and falls on the atmospheric B band. After correction for this absorption, our spectrum gives $\lambda 6583/{\rm H}\alpha$ = 0.80, a normal value for a Seyfert 2 galaxy.

NGC 3185 is a SBa galaxy (Sandage & Bedke 1994). Its emission-line spectrum is power-law photoionized according to Stauffer (1982). Ho et al. (1997a) called it a Seyfert 2, although the published line ratios indicate a "transition'' spectrum. We have fitted the lines with two sets of Gaussian profiles: one system is a H II region; the other corresponds to a Seyfert nebulosity, although the [O I] lines are quite weak ($\lambda 6300/{\rm H}\alpha$ = 0.04).

Arp 107A is the SW component of an interacting galaxy pair (Arp 1966). It exhibits Seyfert 2 activity and its spectrum shows very strong [N II] lines ($\lambda 6583/{\rm H}\alpha$ = 3.0) according to Keel et al. (1985). Our spectrum shows a more normal value. In fact, the lines have a complex profile which can best be fitted by two sets of Gaussians having different widths, but similar $\lambda 6583/{\rm H}\alpha$ ratios, namely 1.41 and 1.23.

NGC 3504. This Sb galaxy (Sandage & Bedke 1994) has a "composite'' nucleus showing both non-thermal activity and recent star formation, the optical spectrum being dominated by the regions of stellar activity (Keel 1984). Ho et al. (1993a), who published line ratios for this object, suggested that it could be a "transition'' between a H II region and a Liner, but Ho et al. (1997a) called it a H II region. Fitting our red spectrum with a single set of Gaussians does not give a satisfactory fit and indicates the presence of weak broad wings in the [N II] lines; two sets of Gaussians are needed, revealing the presence of a weak Seyfert-like nebulosity with $\lambda 6583/{\rm H}\alpha$ = 1.87. The blue spectrum was also fitted by two sets of Gaussians satisfying, respectively, $\lambda 5007/{\rm H}\beta$ = 0.1 and $\lambda 5007/{\rm H}\beta$ = 10. The NGC 3504 spectrum is therefore "composite'' and dominated by a starburst; a weak Seyfert feature is also present. No component showing Liner characteristics was detected in this object.

IRAS 11058 - 1131. In Paper I we concluded, on the basis of a red spectrum, that this object, classified as a Seyfert 2 by various authors (de Grijp et al. 1992; Osterbrock & de Robertis 1985), has weak [N II] lines. Re-analyzing the red spectrum, we have found the possible presence of a weak broad ($FWHM \sim 2\,100$ $\rm km\, s^{-1}$) $\rm H\alpha$ component with 24% of the total $\rm H\alpha$ flux. Our blue spectrum confirms that this is indeed an AGN with $\lambda 5007/$/ $\rm H\beta = 7.6$ and a relatively strong He II $\lambda$4686 line ($\lambda 4686/{\rm H}\beta$ = 0.18). Our conclusion is that IRAS 11058-1131 belongs to the relatively rare class of Seyfert 2 galaxies having weak [N II] lines, discussed in Sect. 4.3; other galaxies belonging to this class are UM 85 and 3C 184.1 (Paper I).

NGC 3642 is an Sb galaxy (Sandage & Bedke 1994). It was classified as a Liner by Heckman (1980). Filippenko & Sargent (1985) noted the presence of very narrow emission lines ($FWHM \sim 110$ $\rm km\, s^{-1}$) superposed on weak, significantly broader components; a weak broad $\rm H\alpha$ component was also detected. Koratkar et al. (1995) confirmed the presence of the broad $\rm H\alpha$ component and called this object a Liner, although their published line ratios rather point to a "transition'' spectrum. Our spectra confirm the presence of a broad $\rm H\alpha$ component ($FWHM \sim 2\,160$ $\rm km\, s^{-1}$) and show, in addition, that the narrow line spectrum is made of two clouds, one with line ratios typical of a H II region and the other of a Seyfert 2 cloud, this object being, therefore, a "composite''.

Mark 1291. Spectroscopic observations of this barred spiral by Kollatschny et al. (1983) show it to be a "transition object'' between Seyfert 2s and H II regions; however, high excitation lines characteristic of Seyfert 2s such as [Fe XIV]$\lambda$5303 and [Fe X]$\lambda$6374 are detected. The nuclear emission is compact (González Delgado et al. 1997). Our optical spectra show a weak broad ($FWHM \sim 1\,950$ $\rm km\, s^{-1}$) $\rm H\alpha$ component suspected by Kollatschny et al. The narrow lines are well fitted with a single Gaussian profile ($\sim 160$ $\rm km\, s^{-1}$ FWHM). The line ratios are $\lambda 5007/{\rm H}\beta$ = 3.84, $\lambda 6300/{\rm H}\alpha$ = 0.04 and $\lambda$6583/ $\rm H\alpha$ = 0.73 (significantly larger than the published value, 0.48); the classification of this object is therefore ambiguous: it is a Seyfert 2 in the the $\lambda 5007/{\rm H}\beta$ vs. $\lambda 6583/{\rm H}\alpha$ diagram and a H II region in the $\lambda 5007/{\rm H}\beta$ vs. $\lambda 6300/{\rm H}\alpha$ diagram. It would be of interest to obtain high-resolution, high signal-to-noise spectra of this object to confirm its "transition'' nature.

IRAS 11285 + 8240 A has been classified as a Seyfert 2 galaxy by Klaas & Elsasser (1991), with $\lambda 5007/{\rm H}\beta$ = 8.25 and $\lambda 6583/{\rm H}\alpha$ = 0.46. Our red spectrum is well fitted by a single set of components with $\lambda 6583/{\rm H}\alpha$ = 0.45, in excellent agreement with the published value, and $\lambda 6300/$/ $\rm H\alpha = 0.11$. This is another example of Seyfert 2 galaxy with marginally weak [N II] lines.

Mark 739 has a double nucleus (Petrosian et al. 1978). The eastern nucleus has a Seyfert 1 spectrum (Petrosian et al. 1979; Netzer et al. 1987). The western component, Mark 739W, has a starburst spectrum according to Netzer et al.; however, Rafanelli et al. (1993) have remarked that $\rm H\beta$ is unresolved with a resolution of 130 $\rm km\, s^{-1}$, while the [O III] lines are significantly broader ($\sim 300$ $\rm km\, s^{-1}$ FWHM). Our observations show that this spectrum is indeed "composite'' with a Seyfert 2 nucleus and a H II region and, in addition, a weak broad $\rm H\alpha$ component.

SBS 1136 + 594 is a Seyfert 1.5 galaxy (Markarian et al. 1983; Martel & Osterbrock 1994). The narrow-line spectrum, however, has very weak [N II] lines ($\lambda 6583/$ $\rm H\alpha$ = 0.25; Martel & Osterbrock); this is confirmed by our spectra, with even weaker [N II] lines ($\lambda 6583/$/$\rm H\alpha = 0.10$). This object belongs to the class of AGNs with very weak [N II] lines discussed in Sect. 4.3. The individual components used to fit the red spectrum of this object are plotted in Fig. 1.

 

Figure 1:  Blue and red spectra for the 53 galaxies studied in this paper. For 52 of the observed AGNs we present 66 Åmm-1 spectra; in the case of VII Zw 73, we give 33 Åmm-1 spectra. For 4 of the observed objects, only one spectral region is available. All the spectra were de-redshifted to rest wavelengths. The spectral ranges displayed are $\lambda\lambda\,4750-5120$ Å and $\lambda\lambda\,6480-6650$ Å. In each frame the data points (small crosses), the best fit (solid line) and the residuals (lower solid line) are shown. For the red spectrum of SBS 1136+594, the individual components of the fit are also given as an example (dotted lines)

 

Figure 1: continued

 

Figure 1: continued

 

Figure 1: continued

 

Figure 1: continued

 

Figure 2:  $\rm H\alpha$+[N II] HST image of IRAS 04210+0400, after removal of the continuum. The nucleus is indicated by an arrow. 10$^{\prime\prime}$, given as a reference on the image, correspond to 13.5 kpc at the distance of the galaxy (assuming H0 = 50 $\rm km\, s^{-1}$ Mpc-1)

NGC 3994 is an Sbc galaxy (Sandage & Bedke 1994) in a triple system, interacting both with NGC 3991 and NGC 3995. Based on the observed line ratios, Keel et al. (1985) classified it as a Liner; the $\lambda 5007/{\rm H}\beta$ and $\lambda 6583/{\rm H}\alpha$ values rather suggest a H II region. We found that its spectrum is "composite'', the main contribution to the Balmer lines coming from a starburst. The relative strength of the [O I] lines is large in the other component ($\lambda 6300/{\rm H}\alpha$ = 0.53), suggesting that it is a Liner.

NGC 4102 is a Sb galaxy (Sandage & Bedke 1994). For Ho et al. (1997a) it is a H II region, although its UV spectrum does not resemble that of a starburst galaxy (Kinney et al. 1993). The spectrum published by Ho et al. (1995) rather indicates a "transition object''. Our blue spectrum shows a $\rm H\beta$ line obviously narrower than the [O III] lines, indicating that the spectrum is "composite''. The blue spectrum was fitted with two sets of three Gaussians. The broadest $\rm H\beta$ profile in the fit contains 7% of the total $\rm H\beta$ flux. The red spectrum having a very high signal-to-noise ratio, we needed three sets of three Gaussians to get a good fit; we forced one set to have the same width as the broadest set in the blue fit. The fitting analysis gives two strong narrow components and a weak broad one, containing $\sim$ 5% of the total $\rm H\alpha$ flux and having $\lambda 6583/{\rm H}\alpha$ = 1.57. There was no need to use two narrow components to fit the $\rm H\beta$ line, as the spectrum was taken under poor transparency conditions and its signal-to-noise ratio is much lower. Our conclusion is that the nucleus of NGC 4102 is dominated by a starburst, but that a weak Seyfert 2 component is present and detected mainly by the broadening of the [O III] lines.

IRAS 12474 + 4345 S. For de Grijp et al. (1992), this object is a H II region; however, the published $\lambda 5007/{\rm H}\beta$ line ratio (2.93) is slightly high for the corresponding $\lambda 6583/{\rm H}\alpha$ ratio (0.42). Fitting our spectra with a single set of lines gives an unsatisfactory result, while the solution with two sets of lines is acceptable, with small residuals. One set of lines corresponds to a H II region, while the other has $\lambda 5007/{\rm H}\beta$ = 4.43, $\lambda 6300/{\rm H}\alpha$ = 0.03 and $\lambda 6583/{\rm H}\alpha$ = 0.40; these values are intermediate between those corresponding to H II regions and Seyfert 2 nebulosities. Moreover, the He II$\lambda$4686 line is detected with $\lambda 4686/{\rm H}\beta$ = 0.17, if all the He II flux is attributed to the "transition'' component. We are unable to conclude concerning the nature of this second component.

UGC 8621 is a Seyfert 1.8 galaxy according to Osterbrock & Martel (1993). Our blue spectrum shows the $\rm H\beta$ line to be much narrower (< 80 $\rm km\, s^{-1}$) than the [O III] lines ($\sim$ 665 $\rm km\, s^{-1}$), but no evidence of a broad component. To account for the different widths observed, we fitted the blue spectrum with two sets of lines; for the narrow component, we found $\lambda 5007/{\rm H}\beta$ = 0.28, while we have forced the broader component to have $\lambda 5007/{\rm H}\beta$ = 10. We fitted the red spectrum with two sets of three Gaussians, imposing to one of them to have the same width as the broader [O III] component; an additional Gaussian was added to fit the broad $\rm H\alpha$ wings. We find no trace of [O I] emission ($\lambda 6300/{\rm H}\alpha$ $\leq$ 0.01) for the narrow component and $\lambda 6300/{\rm H}\alpha$ < 0.3 for the Seyfert cloud. This is a "composite object'', with a relatively strong H II region and a weak Seyfert 1.9 nebulosity.

Mark 266 is a merging system with two nuclei separated by 10$^{\prime\prime}$ (Hutchings & Neff 1988; Wang et al. 1997). It is a Luminous Infrared Galaxy (LIG), i.e., 10^11.2 $< L\rm_{IR} <$ 10¹² $L_{\hbox{$\odot$}}$ (Goldader et al. 1997). Line intensity ratios in the two nuclei have been measured by Osterbrock & Dahari (1983); Veilleux & Osterbrock (1987); Mazzarella & Boroson (1993); Osterbrock & Martel (1993) and Kim et al. (1995); all these measurements are in good agreement, if we make the assumption that Kim et al. have inverted the NE and SW components. On the basis of these line ratios, Mark 266SW has been classified as a Seyfert 2 by Mazzarella & Boroson, Kim et al. and Wu et al. (1998b), and as a "marginal'' Seyfert 2 by Osterbrock & Dahari, while Mark 266NE has been called a Liner by Mazzarella & Boroson, Kim et al. and Wu et al., and a "narrow emission-line galaxy'' (NELG) by Osterbrock & Dahari and Veilleux & Osterbrock. NELGs, for these authors, are emission-line galaxies that may be either Liners or H II regions. For Mark 266NE, we have only a blue spectrum which, by simple visual inspection, shows quite different $\rm H\beta$ and [O III] line profiles. Our line profile analysis reveals two clouds, one with narrow lines ($FWHM \sim 300$ $\rm km\, s^{-1}$) and $\lambda 5007/{\rm H}\beta$ = 0.96, the other with broader lines ($FWHM \sim 1\,000$ $\rm km\, s^{-1}$) and $\lambda 5007/{\rm H}\beta$ = 2.2 and containing 37% of the total $\rm H\beta$ flux. The published value of the intensity ratio $\lambda 6300/{\rm H}\alpha$ is 0.15. If the narrow component is associated with a H II region, it contributes in a small amount to the observed [O I]$\lambda$6300 flux and therefore the $\lambda 6300/{\rm H}\alpha$ ratio for the broader component is likely to be significantly larger than 0.12, which means that this component could be a Liner. We have fitted the blue spectrum of Mark 266SW with three sets of Gaussians: one of them corresponds, most probably, to a H II region with narrow lines (295 $\rm km\, s^{-1}$ FWHM) and $\lambda 5007/{\rm H}\beta$ = 0.36; the two other sets have widths of 200 and 600 $\rm km\, s^{-1}$ and $\lambda 5007/{\rm H}\beta$ = 3.7 and 13.3, respectively. We also fitted the red spectrum with three sets of Gaussians, forcing, in addition, one of the sets to have a width of 600 $\rm km\, s^{-1}$. The result is a set of narrow lines with $\lambda 6583/{\rm H}\alpha$ = 0.58 corresponding to the narrow blue lines, confirming that this system is indeed coming from a H II region. The set having the broadest lines has intensity ratios typical of a Seyfert cloud. The third set, with $\lambda 5007/{\rm H}\beta$ = 3.65 and $\lambda 6583/{\rm H}\alpha$ = 0.58, has still an intermediate spectrum.

Mark 1361 was called a Seyfert 2 galaxy by Kim et al. (1995). Our analysis of a red spectrum (Paper I) lead to the conclusion that it is a "composite object''. Our blue spectrum confirms this result. Three sets of three components were needed to get a good fit. In one set we had to impose $\lambda 5007/{\rm H}\beta$ = 10, in another $\lambda 5007/$ $\rm H\beta$ = 0.1. The best fit resulted in a narrow set of lines with very weak [O III] lines and two sets of lines with strong [O III] contribution. We then re-analyzed the red spectrum using three sets of three lines; we had to add a weak broad $\rm H\alpha$ component ($FWHM \sim 2\,400$ $\rm km\, s^{-1}$, with $\sim$ 10% of the total $\rm H\alpha$ flux) in order to obtain a good fit. For the narrowest set of three lines, we found $\lambda 6583/{\rm H}\alpha$ = 0.54, for the other two, 0.66 and 1.04 respectively. The conclusion is that Mark 1361 has a "composite'' spectrum with a H II cloud contributing half of the $\rm H\beta$ line and a Seyfert 2 nebulosity with complex line profiles (two Gaussians were needed for the fit). If the presence of the broad $\rm H\alpha$ component is confirmed, this object could be a Seyfert 1.9 galaxy.

Mark 461 is a Seyfert 2 galaxy according to Huchra & Burg (1992); however, Cruz-González et al. (1994) have measured $\lambda 5007/{\rm H}\beta$ = 1.13. The emission is concentrated in the nuclear region (González Delgado et al. 1997). No other line ratios have been published for this object. The $\rm H\beta$ and [O III] lines obviously do not have the same profile. To get a good fit, two sets of lines were necessary. The object is "composite'', one component being a H II region and the other probably a Seyfert 2 nucleus.

4C 26.42. This object has been identified with a cD galaxy, MCG 05.33.005 (Carswell & Wills 1967; Olsen 1970; Merkelijn 1972), the brightest member of Abell 1975 (Parma et al. 1986; Pilkington 1964); it is a FR I, Z-shaped radiosource (van Breugel et al. 1984; Ge & Owen 1993). Emission lines have been detected in the nuclear region, with $\lambda 5007/{\rm H}\beta$ = 0.4, $\lambda 6583/{\rm H}\alpha$ = 0.8 and $\lambda 6300/{\rm H}\alpha$ = 0.2 (Anton 1993). These values are similar to the ones usually observed in Liners, but for the low [O III]$\lambda$5007 line intensity. Examination of the red spectrum shows that the lines are obviously double. Fitting the lines with two sets of components revealed two clouds with a velocity difference of $\sim$ 330 $\rm km\, s^{-1}$. Their line ratios are very similar and typical of Liners with exceptionally weak [O III] lines.

I Zw 81. Koski (1978) observed narrow ($FWHM = 225 \pm 200$ $\rm km\, s^{-1}$) emission lines in this galaxy, with $\lambda$5007/$\rm H\beta$ = 3.14, $\lambda$6583/$\rm H\alpha$ = 0.67 and $\lambda 6300/{\rm H}\alpha$ = 0.07; the narrowness of the emission lines and the relative weakness of the [O III] lines led him to conclude that this is not a Seyfert galaxy, but rather a "transition'' case between a H II galaxy and a Seyfert 2. Shuder & Osterbrock (1981) and Veilleux & Osterbrock (1987) called it a narrow-line Seyfert. Our red spectrum shows the lines to be double with a separation of 220 $\rm km\, s^{-1}$. Fitting both the red and blue spectra with two sets of three Gaussians, we found $\lambda 5007/{\rm H}\beta$ = 2.05 (1.41) and $\lambda 6583/{\rm H}\alpha$ = 0.78 (0.56) for the high (low) velocity clouds. In both cases, the [O I] lines are undetected with $\lambda 6300/{\rm H}\alpha$ < 0.04. The two clouds are most probably H II regions.

Mark 477. This object was discussed in Paper I, where we concluded from the published data and the analysis of a red spectrum that its nature was unclear. Heckman et al. (1997) have argued that the observed UV through near-IR continuum in the nucleus of Mark 477 is dominated by light from a starburst. The narrow Balmer emission lines would then be excited by ionizing radiation from both the hidden Seyfert 1 nucleus and from the hot stars in the starburst. We have re-analyzed our old red spectrum together with our new blue spectrum. Three sets of components were needed to fit these very high signal-to-noise data. The three line-systems are typical of Seyfert 2s, although two have relatively weak [N II] intensities, with $\lambda$6583/ $\rm H\alpha$ = 0.31 and 0.38, respectively. We found no evidence for the presence of broad Balmer components.

Mark 848S is a LIG (Goldader et al. 1997) belonging to a pair of interacting galaxies (Armus et al. 1990). The northern galaxy is a H II region (Wu et al. 1998a,b); the southern component has been called a Liner (Mazzarella & Boroson 1993; Veilleux et al. 1995), though its line ratios (Kim et al. 1995) are ambiguous, the [O I]$\lambda$6300 line being rather weak for a Liner. The line profile analysis of our spectra shows it to be a "composite object'' with two distinct emission-line clouds: a narrow one ($FWHM \sim 140$ $\rm km\, s^{-1}$, with $\lambda 5007/{\rm H}\beta$ = 0.83, $\lambda 6583/{\rm H}\alpha$ = 0.43 and $\lambda 6300/{\rm H}\alpha$ = 0.03), identified with a H II region, and a somewhat broader one ($FWHM \sim 580$ $\rm km\, s^{-1}$), with line ratios typical of a Seyfert 2 ($\lambda 5007/{\rm H}\beta$ = 4.20, $\lambda$6583/ $\rm H\alpha$ = 0.71 and $\lambda 6300/{\rm H}\alpha$ = 0.14).

IRAS 15184 + 0834. De Grijp et al. (1992) called this object a Seyfert 2; however they found relatively weak [N II] lines ($\lambda 6583/{\rm H}\alpha$ = 0.42). Our spectra show that the [N II] lines are significantly stronger than the published values. A reasonable fit is obtained with two sets of components: one is a H II region; the other could be a Seyfert 2, although the [O I] lines are relatively weak.

NGC 5953 is a peculiar S0 (Rampazzo et al. 1995) or Sa (González Delgado & Pérez 1996) galaxy interacting with NGC 5954 (Arp 1966). It has a Seyfert 2 nucleus (Rafanelli et al. 1990; González Delgado & Pérez) surrounded by a ring of star formation with a radius of $\sim$ 4$^{\prime\prime}$ (González Delgado & Pérez). Rafanelli et al. and González Delgado & Pérez studied this object using a slit width of 2$.\!\!^{\prime\prime}$0 and 1$.\!\!^{\prime\prime}$5 respectively. The seeing was $\sim$ 1$^{\prime\prime}$ during González Delgado & Pérez' observations, while it was not specified by Rafanelli et al. who, however, easily separated the galaxy nucleus from a star located 3$^{\prime\prime}$ away. We may reasonably assume that, in both cases, the nuclear spectrum corresponds to a 2$.\!\!^{\prime\prime}$0 $\times$ 2$.\!\!^{\prime\prime}$0, or smaller, aperture. The line ratios given by these authors (see Table 2) are typical of a Seyfert 2 region, although [O I]$\lambda$6300 may be somewhat weak for this type of objects; but as stressed before, we cannot exclude some contamination by the circumstellar emission region. Both Keel et al. (1985) and Kim et al. (1995) have observed the NGC 5953 nuclear region with a relatively large aperture: Keel et al. used a $\phi$ = 4$.\!\!^{\prime\prime}$7 circular aperture, while Kim et al. used a long 2$^{\prime\prime}$ $\times$ 14$^{\prime\prime}$ slit. It is clear that these two large aperture spectra must contain a significant contribution from the circumstellar emission region and, indeed, the published line intensity ratios are those of "transition'' spectra. We used a 2$.\!\!^{\prime\prime}$1 slit and the seeing was 2$.\!\!^{\prime\prime}$6; we extracted 7 columns, i.e., 7$.\!\!^{\prime\prime}$6, so some contamination from the circumstellar emission region was expected. Effectively, by doing the line profile fitting analysis, we were able to identify two components of different line widths and strengths, one of which can be associated with a H II region ($\lambda 5007/{\rm H}\beta$ = 0.55, $\lambda 6583/$/ $\rm H\alpha = 0.60$, $\lambda 6300/{\rm H}\alpha$ = 0.03 and $FWHM \sim 200$ $\rm km\, s^{-1}$); the other component, broader ($FWHM \sim 400$ $\rm km\, s^{-1}$), reveals the presence of a Seyfert 2 nebulosity, the measured line intensities being: $\lambda 6583/{\rm H}\alpha$ = 1.96 and $\lambda 6300/{\rm H}\alpha$ = 0.18, with $\lambda 5007/{\rm H}\beta$ fixed to 10. Lines of [Fe III]$\lambda$4659 and [Fe VII]$\lambda$5158 are clearly visible in the blue spectrum. A very weak broad $\rm H\alpha$ line is possibly detected, which would qualify NGC 5953 as a Seyfert 1.9 galaxy.

Kaz 49 has been classified as a Seyfert 1 by Yegiazarian & Khachikian (1988), as a Seyfert 1.9 by Moran et al. (1994), who have detected a weak broad $\rm H\alpha$ component ($FWHM = 1\,150$ $\rm km\, s^{-1}$), and as a H II region by Boller et al. (1994). The latter classification was based on measured line ratios ($\lambda 5007/{\rm H}\beta$ = 2.58, $\lambda 6583/{\rm H}\alpha$ = 0.56, $\lambda 6300/{\rm H}\alpha$ = 0.025) that rather point to a "transition'' spectrum. The line profile analysis of our spectra shows a strong H II region ($\lambda 5007/{\rm H}\beta$ = 2.21, $\lambda 6300/{\rm H}\alpha$ = 0.05, $\lambda 6583/{\rm H}\alpha$ = 0.55) and a weak Seyfert 2 component for which we have fixed $\lambda 5007/{\rm H}\beta$ = 10. There is no evidence for the presence of a broad $\rm H\alpha$ component; however, the blended weak $\rm H\alpha$ and [N II] components, each having a $FWHM \sim 880$ $\rm km\, s^{-1}$ may be easily mistaken for a broad $\rm H\alpha$ line.

IRAS 16129 - 0753 has been classified as a possible Liner by de Grijp et al. (1992) on the basis of the measured line intensity ratios ($\lambda 5007/{\rm H}\beta$ = 2.03, $\lambda 6583/$/ $\rm H\alpha = 0.64$), although [O I]$\lambda$6300 was very weak. The line fitting analysis of our blue spectrum shows this object to be "composite''. The red spectrum, which has a relatively low signal-to-noise ratio, is well fitted by a single set of lines corresponding to the H II region; the Seyfert component is undetected.

IRAS 16382 - 0613 has been called a Seyfert 2 by Aguero et al. (1995) and a possible Seyfert 2 by de Grijp et al. (1992); however, the [O I]$\lambda$6300 line is marginally weak for a Seyfert 2, with $\lambda 6300/{\rm H}\alpha$ = 0.09 (Aguero et al.). The line profiles on the blue spectrum are obviously complex. Fitting these lines with two sets of Gaussians reveals a narrow component ($FWHM \sim 350$ $\rm km\, s^{-1}$) with $\lambda 5007/{\rm H}\beta$ = 3.94, and a broader component ($FWHM \sim 1\,160$ $\rm km\, s^{-1}$) with $\lambda 5007/{\rm H}\beta$ = 4.06. The red spectrum fit gives a solution compatible with the blue solution plus a broad Balmer line ($FWHM \sim 5\,000$ $\rm km\, s^{-1}$). The two components have strong [N II] lines, but the [O I] lines are weak. For the broadest set of lines, we find $\lambda 6300/{\rm H}\alpha$ < 0.12, compatible with a Seyfert 2 nebulosity; however, the narrow component has $\lambda 6300/{\rm H}\alpha$ < 0.03 and seems therefore to have a genuine "transition'' spectrum.

Mark 700 was called a Seyfert 1 galaxy by Denisyuk et al. (1976), who found a broad $\rm H\alpha$ component. For Koski (1978), it is a weak-lined Seyfert galaxy with Balmer absorption lines, very similar to "normal'' emission-line galaxies. Ferland & Netzer (1983) included it in a Liner list, on the basis of the intensity ratios published by Koski. Our observations show that this object is, indeed, a Liner; the broad $\rm H\alpha$ component seen by Denisyuk et al. is confirmed.

MCG 03.45.003. The [N II] lines measured by de Grijp et al. (1992) are rather weak for a Seyfert 2 galaxy ($\lambda 6583/{\rm H}\alpha$ = 0.42) and, on the basis of a red spectrum, we concluded in Paper I that this object could have a "composite'' spectrum. Our analysis of both the blue and red spectra show that two kinematically distinct clouds are present in this object, both of them having Seyfert 2 characteristics.

PGC 61548. The red spectrum is "composite'' and confirms the result presented in Paper I. The line profile analysis reveals the presence of both a H II region ($\lambda$5007/$\rm H\beta$ = 0.41, $\lambda$6583/$\rm H\alpha$ = 0.50, $\lambda$6300/$\rm H\alpha$ = 0.04, $FWHM \sim 250$ $\rm km\, s^{-1}$) and a Seyfert 2 nebulosity ($\lambda 5007/{\rm H}\beta$ fixed to 10.0, $\lambda 6583/{\rm H}\alpha$ $\sim$ 3.9, $\lambda 6300/{\rm H}\alpha$ $\sim$ 0.5 and $FWHM \sim 570$ $\rm km\, s^{-1}$).

Kaz 214 is a Seyfert 2 galaxy for de Grijp et al. (1992), with $\lambda 5007/{\rm H}\beta$ = 5.23 and $\lambda 6583/{\rm H}\alpha$ = 0.39; however, the [N II] lines are weak for a Seyfert 2. On our red exposure, the slit position angle was PA = 139$^\circ$. By simple visual inspection of the CCD image, we see that the lines are double: in one of the line-systems the lines are spatially extended and narrow, with relatively weak [N II]; in the other, the lines are spatially unresolved, but relatively broad, and [N II] is stronger. The spectrum is obviously "composite'' with a H II region and a Seyfert component. However, when analysing the spectrum obtained by extracting three columns centered on the nucleus, we were unable to get a satisfactory fit confirming the visual impression. We then extracted individually seven columns (numbered 1 to 7, from SE to NW) containing obvious emission lines; the continuum was brightest in Cols. 4 and 5. Columns 1, 2 and 7 were fitted with a single set of lines, while for Cols. 3 to 6, two sets of lines were necessary. We have made the assumption that the Seyfert component is really spatially unresolved and, consequently, forced the redshift, width and the $\lambda 6583/{\rm H}\alpha$ ratio of this component to be the same on all columns (that is, 120 $\rm km\, s^{-1}$, 525 $\rm km\, s^{-1}$ FWHM and 0.60, respectively), the only free parameter being the $\rm H\alpha$ intensity. In addition to this Seyfert component, we have found, on all columns, a narrow component with relatively weak [N II] lines; the velocity of this narrow component increases from -25 $\rm km\, s^{-1}$ to 140 $\rm km\, s^{-1}$ from Cols. 1 to 7. The blue spectrum was taken with the slit oriented N-S. As the seeing was rather poor, seven columns were added together when extracting the spectrum. The best fit was obtained with three sets of lines: for one of them, we forced $\lambda 5007/{\rm H}\beta$ = 10 (this turns to be the broadest component); the two other sets have narrow lines, with moderate $\lambda 5007/{\rm H}\beta$ ratios. We therefore conclude that Kaz 214 has a "composite'' spectrum. But this example shows that it may not be possible to show that a "transition'' spectrum is a "composite'' spectrum when the spatial resolution is insufficient, this being due to the large velocity dispersion gradient sometimes present in the nuclear region which broadens the lines.

NGC 6764 has been called a Seyfert 2 galaxy by Rubin et al. (1975), in spite of $\rm H\beta$ being stronger than [O III]$\lambda$5007; this classification was based on the width of the $\rm H\alpha$ and [N II] lines ($\sim$ 750 $\rm km\, s^{-1}$) but Wilson & Nath (1990) found these lines to be much narrower ($\sim 380$ $\rm km\, s^{-1}$ FWHM). Koski (1978) noticed the presence of weak H I absorption lines and, from the line intensity ratios, concluded that it was very much like "normal'' galaxies, while for Shuder & Osterbrock (1981), it is not a Seyfert 2. Using the line ratios published by Koski (1978); Ferland & Netzer (1983) classified it as a Liner. Osterbrock & Cohen (1982) have detected in the spectra of this object the $\lambda$4650 Wolf-Rayet emission feature. For Ashby et al. (1992), it is a starburst galaxy. Line profile fitting of our spectra revealed the "composite'' nature of this object. Two systems were identified: a narrow one (FWHM = 325 $\rm km\, s^{-1}$) with line ratios compatible with those usually observed in H II regions ($\lambda 5007/{\rm H}\beta$ = 0.62, $\lambda 6583/{\rm H}\alpha$ = 0.65 and $\lambda 6300/{\rm H}\alpha$ = 0.04) and a broader system (FWHM = 480 $\rm km\, s^{-1}$) with line ratios similar to those of Liners ($\lambda 5007/{\rm H}\beta$ = 0.44, $\lambda 6583/{\rm H}\alpha$ = 0.96 and $\lambda 6300/{\rm H}\alpha$ = 0.14). This is, therefore, a "composite object''.

IRAS 22114 - 1109 was classified a Seyfert 2 by Veilleux et al. (1995); however, the [O III] lines are relatively weak for this type of objects ($\lambda 5007/{\rm H}\beta$ = 4.22; Kim et al. 1995). A line profile analysis was performed on the red and blue spectra. The measured line intensities and widths are compatible with the simultaneous presence on the slit of both a H II region ($\lambda 5007/{\rm H}\beta$ = 1.33, $\lambda 6583/{\rm H}\alpha$ = 0.70, $\lambda 6300/{\rm H}\alpha$ < 0.07 and FWHM = 185 $\rm km\, s^{-1}$) and a Seyfert 2 nebulosity ($\lambda 5007/{\rm H}\beta$ fixed to 10.0, $\lambda 6583/{\rm H}\alpha$ = 0.60, $\lambda 6300/{\rm H}\alpha$ = 0.12 and FWHM = 415 $\rm km\, s^{-1}$), so this is another example of a "composite-spectrum object''.

Mark 308 was called a Seyfert 2 galaxy by Popov & Khachikian (1980) and Zamorano et al. (1994); Véron-Cetty & Véron (1986a) classified it as a H II region, although the published line ratios (Table 2) are unlikely for either classes. The analysis of our blue spectrum (Paper I) showed this object to be "composite'' with one narrow component with weak [O III] lines and two broader components with strong [O III] lines; the analysis of our red spectrum confirms this result, the narrow system ($FWHM \sim 155$ $\rm km\, s^{-1}$) being typical of a H II region (with $\lambda 5007/{\rm H}\beta$ fixed to 0.1, $\lambda 6583/{\rm H}\alpha$ = 0.30 and $\lambda 6300/{\rm H}\alpha$ = 0.06) and the two broader line-sets ($FWHM{\rm s} \sim 325$ and 1045 $\rm km\, s^{-1}$, respectively) of Seyfert-like clouds. Moreover, we have detected a weak broad (1725 $\rm km\, s^{-1}$ FWHM) $\rm H\alpha$ component containing $\sim 7\%$ of the total $\rm H\alpha$ flux. The companion galaxy, KUG 2239+200A, at z = 0.024 (Keel & van Soest 1992) and located 53$^{\prime\prime}$ NE of Mark 308, has a H II-like emission-line spectrum.

Mark 522 is a Seyfert 2 galaxy according to Veilleux & Osterbrock (1987); however, the [O III] and [O I] lines are relatively weak ($\lambda 5007/{\rm H}\beta$ = 3.23, $\lambda 6300/{\rm H}\alpha$ = 0.07). Our observations show this object to be "composite'', with two different line systems: one, "narrow'' ($FWHM \sim 100$ $\rm km\, s^{-1}$), typical of a H II region ($\lambda 5007/{\rm H}\beta$ = 0.63, $\lambda 6583/{\rm H}\alpha$ = 0.53 and $\lambda 6300/{\rm H}\alpha$ < 0.04), the other, somewhat broader ($FWHM \sim 240$ $\rm km\, s^{-1}$), associated with a Seyfert 2 nebulosity ($\lambda 5007/$$\rm H\beta$ = 7.87, $\lambda 6583/{\rm H}\alpha$ = 1.50 and $\lambda 6300/{\rm H}\alpha$ < 0.2).

Mark 313 is a Seyfert 2 galaxy according to Osterbrock & Pogge (1987) (with $\lambda 5007/{\rm H}\beta$ = 3.52, $\lambda 6583/{\rm H}\alpha$ = 0.52 and $\lambda 6300/{\rm H}\alpha$ = 0.10), and Moran et al. (1996); the [O III] lines are relatively weak for a Seyfert 2. Images in [O III] and $\rm H\alpha$+[N II] show a very complex structure, with high excitation gas restricted to a symmetric, linear feature (Mulchaey et al. 1996). From a two-component Gaussian fitting of high-dispersion spectra of the nucleus of this object, Maehara & Noguchi (1988) concluded that it is a "composite object'' with a H II region and a Liner nebulosity. Line profile analysis of our spectra reveals the contribution of two different line-emitting regions: one, with $\lambda 5007/{\rm H}\beta$ = 2.29, $\lambda 6583/{\rm H}\alpha$ = 0.44, $\lambda 6300/$$\rm H\alpha$ = 0.10 and narrow width (135 $\rm km\, s^{-1}$ FWHM), is typical of a H II region; the other, much weaker, is not detected in the blue and its line ratios are $\lambda 6583/{\rm H}\alpha$ = 0.71 and $\lambda 6300/{\rm H}\alpha$ = 0.28; it could be either a Seyfert 2 or a Liner, depending on the $\lambda 5007/{\rm H}\beta$ ratio.

Zw 453.062 is a LIG (Goldader et al. 1997); it was called a Liner by Veilleux et al. (1995) on the basis of the measured emission line ratios, while Wu et al. (1998a,b) found its properties to be intermediate between H II regions and Liners, although the [O I] lines are very weak. Our spectra suggest that it is a "composite object'', one component being a Seyfert 2 nebulosity and the other a H II region.

IC 5298 is a LIG (Goldader et al. 1997). Wu et al. (1998a,b) found that its properties are intermediate between H II regions and Liners; it was classified as a Seyfert 2 by Veilleux et al. (1995) although the [O I] lines are rather weak ($\lambda 6300/{\rm H}\alpha$ = 0.05). Our observations suggest that the spectrum is "composite'', being dominated by a H II region.