3 Differences in the aspect of the molecular gas distribution between type 1 and type 2 Seyferts

In this section we employ a model of the molecular gas in order to try and determine the orientation of the molecular gas for each of the galaxies of Curran et al. (2000). The model, which is discussed in more detail in Curran (1998); Curran (2000), provides a very satisfactory fit to the observed data in the Circinus galaxy, a type 2 Seyfert, which agrees closely with other observations of the gas in this galaxy (Veilleux & Bland-Hawthorn 1997; Elmouttie et al. 1998a; Elmouttie et al. 1998b). Due to the close proximity of Circinus we could map the CO emission with the 22'' beam at SEST, but since these galaxies are much further away, we have no such luxury (especially with the eight Southern galaxies). However, since the optically thick molecular gas is unresolved (e.g. $\approx13''$ in the relatively near-by NGC 1365, Sandqvist 1999, cf. the 45'' HPBW for CO $1\rightarrow 0$ at SEST), the observed profiles will be a result of the gas distribution convolved over the velocity range. This technique has previously been used by Downes & Solomon (1998) in Mrk 231. In Figs. 2 and 3 we show the unresolved central 10'' of Circinus in the CO $2\rightarrow 1$ and $1\rightarrow 0$ transitions compared with the model of Curran et al. (1998).

Figure 2: The Circinus model "observed'' with a 22'' (CO $2\rightarrow 1$ at SEST) beam superimposed upon the CO $2\rightarrow 1$ observed spectrum. Taken from Curran et al. (1998)

Figure 3: The Circinus model observed with a 45'' (CO $1\rightarrow 0$ at SEST) beam superimposed upon the CO $1\rightarrow 0$ observed spectrum. Perhaps due to a small pointing error ( ), the observed spectrum is not as symmetrical as that of CO $2\rightarrow 1$ (Fig. 2), although it is seen that the model still provides a good match

The molecular gas, which is assumed to be in a disc distribution, is modelled as:

1. The basic ring model of Curran et al. (1998);
2. The filled ring model in which the intensity at the inner edge of the ring is extrapolated to zero radius, while maintaining a velocity of 100 km s^-1 (Sect. 3.2.3 of Curran et al. 1998);
3. A disc constructed from the filled ring model but with the central velocity extrapolated to zero (Fig. 9 of Curran et al. 1998);
4. The ring+outflow model of Curran et al. (1999). In addition to an outflow perpendicular to the ring, we constructed versions in which an edge-on, a face-on and an outflow of intermediate orientation were added to rings of various inclinations.

Figure 4: The observed spectra of Curran et al. (2000) shown to a velocity resolution of 10 km s-1, except NGC 0034, NGC 1667, UGC 03374, NGC 4593, Mrk 231, NGCs 5347 and 7172 to 20 km s-1 and Mrk 273 (shown relative to z=0.037780) to 40 km s-1. The intensity scale is $T_{\rm A}^*$ and the velocity scale is relative to the local standard of rest (l.s.r.)

These models were then tested at various inclinations and compared to the observed CO spectra (Fig. 4) in order to see if any obvious matches could be found. This was done for the sources in which a clean profile of the CO was obtained; we have excluded the relatively poor NGC 0931, and Mrk 10 detections (Curran et al. 2000). By obvious matches we mean that it was our priority to try to reproduce the observed profile shapes before trying to match the integrated intensities (i.e. velocity widths and antenna temperatures). For this purpose matching the model and observed profile by eye was sufficient since we didn't really expect to constrain the inclination to closer than $\approx20^{\circ}$ accuracy, and as seen from Fig. 5, differences in the inclination are easily discernable. Naturally some profiles are not unique, e.g. for NGC 5135 (Sect. 3.11), where a disc or a ring model may apply. The model was good enough, however, to give an approximate inclination for a particular model and any variations on the disc distribution which fitted the observed profile have also been noted. If reasonable matches were found, the models were then "observed'' with the required HPBW in order to simulate the model at different apparent sizes. As previously (Curran et al. 1999), we (initially) scaled the beam according to optical major axis of the galaxy rather than the distance (velocity). This method assumes that the extent of the ring/disc molecular gas distribution is proportional to the size of the galaxy as opposed to being the same size for all of the sample (i.e. a radius of $\sim600$ pc, Curran et al. 1998). As long as we are not simulating near-by systems, where the beam is filled, Curran et al. 1999), the choice of beam size will only affect the model intensity and not the profile shape. Examples of model profiles are shown in Fig. 5, and a more extensive range of these is shown in Curran (2000).

Figure 5: Examples of various models (further examples are shown in Curran 2000). Left; a disc model, centre; a ring model and right; a ring+(perpendicular) outflow model. The model inclinations and HPBWs used to "observe'' them are shown in each case; a 22'' beam corresponds to Circinus for CO $2\rightarrow 1$ at SEST

3.1 NGC 0034

The observed $2\rightarrow 1$ profile of NGC 0034 generally resembles that of a highly inclined disc/ring model, but with "ears'' at around 5700 and 5900 km s^-1. These in conjuction with the central hump gives the observed spectrum a slightly 3-peaked profile, and its presence is consistent with a ring+outflow model at inclinations of $\approx 60^{\circ }$. The outflow angle is more difficult to constrain, although we can say that the outflow has an inclination of between $0^{\circ}$ (edge-on) and $60^{\circ }$ (normal to the ring), Fig. 6.

Figure 6: A $60^{\circ }$ inclination ring(+outflow) "observed'' with a 175'' beam. The labels refer to the outflow inclination

To summarise, since a ring+outflow gives such a 3-peaked profile (the ring and filled-ring models give a 2-peaked profile), we suggest that the gas may be distributed in such a fashion, although, due to the model limits and the noisier $1\rightarrow 0$ spectrum, all we can say with some confidence is that the gas appears to be in a ring of fairly high inclination ( ). It should be noted that a beam-width of 70'', which is scaled for the optical disc, gives an intensity which is $\approx5$ times the observed. In order to achieve the same intensity, the model has to be "observed'' with a 175'' beam, Fig. 6. Since, assuming a similar small-scale filling to the gas in Circinus, a $\left(\frac{\rm 79~Mpc}{\rm 4~Mpc}\right)\times22''\approx430''$ beam would be required to scale NGC 0034 according to the distance, we conclude that the gas extends 430''/175''=2.5times as far as in Circinus, i.e. to $\approx1.5$ kpc.

3.2 NGC 1068

The width of the $1\rightarrow 0$ line is suggestive of an inclined disc/ring system with the sheer of the profile edges favouring the (filled or unfilled) ring model. Although the profile is too complex to make anything but broad estimates, the profile shape and width suggest that ring has an intermediate inclination ( $40^{\circ }$ $-60^{\circ}$) with the possibility of a nearly edge-on outflow. This is expected to be inclined close to the ionisation cone (Wilson & Tsvetanov 1994; Curran et al. 1999), which should be close to edge-on in a Sy2 nucleus (Antonucci & Miller 1985; Wilson et al. 1988; Tadhunter & Tsvetanov 1989; Wilson & Tsvetanov 1994; Baker & Scoville 1998). Also, although in this case the model is not so certain, the observed intensity is obtained when it is "observed'' with a $\approx70''$ beam. This is twice the width of a beam scaled for the optical extent. Since based upon distance, a beam width of $\approx140''$ is required to observe the model at 15 Mpc, we can estimate the extent of the molecular gas as being twice that in Circinus, i.e. $\approx1.3$ kpc. Thus the model gives similar results to the interferometric results of Tacconi et al. (1994); Papadopoulos (1996) who determine molecular disc inclinations of $30-50^{\circ}$ and $\approx50^{\circ}$(extending to $\approx1.8$ kpc), respectively.

3.3 NGC 1365

Figure 7: The ring only model at $30^{\circ }$ and $40^{\circ }$inclinations observed with a 28'' beam, i.e. NGC 1365 scaled for the apparent optical size when observed in CO $1\rightarrow 0$ at SEST

As seen from Fig. 7, the Circinus (filled and unfilled) ring model gives a very similarly shaped profile to that of the observed for inclinations of $30^{\circ }$ to $40^{\circ }$. In the case of the model, however, the intensity is similar, which suggests that the extent of the gas is nearly proportional to the optical disc size, i.e. nearly 5 times as extended as in Circinus (to a radius of  kpc). From VLA and SEST observations, Sandqvist et al. (1995) find the molecular gas to lie within the central 2.2 kpc.

A significant difference, however, is the fact that the model profile is narrower than the observed by a factor of $\approx2$. This suggests that the maximum de-projected velocity in the ring of NGC 1365 is double that in Circinus i.e. $\approx\pm400$ km s^-1(Curran et al. 1998). Since $M\propto v^2r$, this implies that the molecular gas mass (assuming a relatively small central dynamical mass, e.g. 10% of the gas mass, Curran 1998) is (at least) four times that in Circinus. Interestingly, Sandqvist et al. (1995) find the molecular gas mass in the central region of NGC 1365 to be $6.3\ 10^9 M_{\odot}$, i.e. also 4 times that in Circinus (Curran et al. 1998). It is possible, however, that since the extent of the gas in NGC 1365 may be considerably larger, this factor would require an even larger gas mass/different physical conditions to those in Circinus.

Like Circinus (Marconi et al. 1994; Veilleux & Bland-Hawthorn 1997; Elmouttie et al. 1998b), NGC 1365 exhibits an ionisation cone Hjelm & Lindblad 1996) and since the model cannot distinguish the presence of an outflow of low inclination, we cannot exclude the presence of this (Curran 2000). It should be noted, however, that the observed profile may arise from the presence of the strong bar in NGC 1365 (e.g. Teuben et al. 1986; Lindblad et al. 1996), although, as yet, current observations (e.g. Sandqvist et al. 1995) are of insufficient resolution in order to distinguish whether the CO distribution expected from the bar is present (Kenney et al. 1992; Kenney 1996). In summary, all we can say is that, if a molecular ring is the dominant component on the sub-kpc scale, then our model suggests that it has an inclination of between $30^{\circ }$ and $40^{\circ }$ which puts it coplanar to the obscuring torus and large scale galactic disc (Wilson & Tsvetanov 1994; Hjelm & Lindblad 1996).

3.4 NGC 1667

Since the detection of NGC 1667 may have been compromised by a pointing error, Fig. 4, we assume that this has a symmetric form as in Papadopoulos & Seaquist (1998). Although gentler, the profile has a similar shape to that of NGC 1365. When we observe ring models at inclinations of $\approx30^{\circ}$ with a 125'' beam (CO $1\rightarrow 0$ at OSO) we obtain a similar general shape and intensity, although the velocity range is again about a factor of 2 too small. Since Circinus has a similar CO luminosity within 800 pc as NGC 1667 out to $\approx5$ kpc ( $\approx 1~10^{3}$ K km s^-1 kpc², Curran et al. 1998; Curran et al. 2000), a smaller extent in the molecular ring/disc could account for the higher velocities. Of particular interest, however, is the fact that the addition of a (strong, close to edge-on) molecular outflow is required in order to produce the observed central hump (Fig. 4 and Maiolino et al. 1997; Papadopoulos & Seaquist 1998).

3.5 UGC 03374

The low signal-to-noise (S/N) ratio in this source does not allow us to be quite as specific when fitting a model, but the observed spectrum (Fig. 4 and Maiolino et al. 1997) may be fitted by a disc of any inclination or a ring of inclination . In order to account for the observed intensity, a beam width exceeding 200''(the limit of the model program) is required, although the lower luminosity ( 0.4 10³ K km s^-1 kpc², Curran et al. 2000) could contribute to this.

3.6 NGC 2273

Again the S/N ratio is too low in order to distinguish particular details (i.e. the presence of an outflow), although the general shape does suggest a (filled or unfilled) ring rather than a disc. From the shape of the profile, we estimate this ring to have an inclination of between $20^{\circ }$ and $50^{\circ}$. Again, in order to account for the observed intensity, a beam width exceeding 200'' is required. This is the size of beam required to scale NGC 2273 according to distance, and so we suggest that the molecular gas is more confined than the ring/disc in Circinus, although again the lower CO luminosity ( 0.4 10³ K km s^-1 kpc², Young et al. 1995) may affect this.

3.7 NGC 4593

Again in the case of the CO $1\rightarrow 0$ profile, the S/N ratio is too low to distinguish any details, although the shape of the spectrum does suggest a close to edge-on ring or disc, thus orientating the gas close to perpendicular to the torus of the Sy1 nucleus. Looking at the CO $2\rightarrow 1$ profile (Fig. 4), we see, however, that the width may be due to a 3-component system. Such a profile is what we might expect if a close to face-on ring+outflow were present (Curran et al. 1999). With a face-on (Sy1) outflow, the observed spectrum is best fitted with a ring of inclination, Fig. 8.

Figure 8: The ring (at inclinations of $10^{\circ }$ and $20^{\circ }$)+outflow (close to face-on) model observed with a 200'' beam, i.e. scaled for a 600 pc ring at 39 Mpc. The fact that this is still $\approx 10$ times the observed intensity suggests that the molecular gas extent is much less than in Circinus. The optical disc scale (39'' beam for CO $2\rightarrow 1$ at SEST) gives an intensity of over 10 times that shown (i.e. 100 times the observed). In the figure the outflow features ( $\approx \pm 180$ km s-1) overlap for both ring inclinations

One difference between the model and the observed spectrum is the fact that the peripheral (model outflow) features occur at $\approx\pm90$ km s^-1 in the observed spectrum. This "slower outflow'', cf. Circinus, may be due to the fact that NGC 4593 is considerably less luminous (Curran et al. 2000) and this would also explain the low observed intensity, in addition/as an alternative to a confined molecular gas distribution, Fig. 8.

3.8 Mrk 231

Although the S/N ratio is significantly poorer, like NGC 5135 (Sect. 3.11) the observed $1\rightarrow 0$  profile of Mrk 231 could be due to either a disc or ring distribution at fairly low ( ) inclinations. Note that Downes & Solomon (1998) find a nearly face-on molecular disc from their model. From the intensity of the observed profile we estimate an extent of CO emission which is about 7 times that in Circinus, i.e. out to $\sim4$ kpc. This is an order of magnitude higher than the radius determined from the interferometric measurements of Bryant & Scoville (1996), and in this case we must therefore conclude that the relatively high observed intensity is a result of the high intrinsic CO luminosity in Mrk 231 ( times Circinus within the beam, Curran et al. 1998; Curran et al. 2000) rather than due to a large extent. This emphasises that any values we obtain for the CO extent, should be considered as limits only, Table 2.

3.9 NGC 5033

As with CO  $2\rightarrow 1$ in NGC 0034, the spectra of Maiolino et al. (1997); Papadopoulos & Seaquist (1998) suggest a (filled or unfilled) ring, possibly with an outflow, which has a fairly high ( ) inclination, although this is not so apparent in our noisy $1\rightarrow 0$detection, Fig. 4. The beam size used to obtain the observed intensity is based upon distance. Although as in the case of Mrk 231 (Sect. 3.8), this may be a consequence of the high intrinsic luminosity (Young et al. 1995).

3.10 Mrk 273

In Mrk 273 the S/N ratio is too poor in order to assign a model. From the intensity, however, we can estimate that a HPBW of 400''produces the observed intensity, as opposed to the 1200'' beam expected if the molecular gas in Circinus were scaled for distance. This leads us to conclude that the molecular gas extends to around $\sim2$ kpc. This is comparable with the $\approx2.5$ kpc obtained from the interferometric map of Yun & Scoville (1995). Assuming a flat circular distribution, the disc has an inclination of $\approx 60^{\circ }$, which we shall adopt.

3.11 NGC 5135

Both the $1\rightarrow 0$ and CO  $2\rightarrow 1$ observed spectra can only be fitted with disc/ring model of low inclination, although this produces a somewhat narrow profile (Fig. 9). As before (Sect. 3.4), a more confined gas could account for the high velocities, although the addition of an outflow could achieve this, Fig. 9. A sufficiently high S/N ratio for the profile wings would be required in order to see this in the observed profile, Fig. 4.

Figure 9: The disc and ring ( $10^{\circ }$)+outflow (edge-on) model observed with a 122'' beam, i.e. scaled for NGC 5135 observed in CO $1\rightarrow 0$ at SEST. The fact that the model intensity is similar to that of the observed, suggests that the molecular gas does scale with the optical disc for this luminous galaxy. Since we would require a 630'' beam to place NGC 5135 at the same distance as Circinus we estimate that the ring/disc molecular gas is approximately 5 times as extended. This contradicts the fact that the profile width suggests a more confined gas and so we attribute the observed width to the higher intrinsic luminosity of NGC 5135 within the beam (Curran et al. 2000)

3.12 NGC 5347

When fitting the model, the narrow profile of this source is reminiscent of a lowly inclined disc or ring system (of $\sim600$ pc extent). It should be noted, though, that NGC 5347 has a much lower luminosity than Circinus and so a dimmer (i.e. slower) system of higher inclination is possible.

3.13 NGC 5548

The $2\rightarrow 1$ (and possibly $1\rightarrow 0$) detection of this is reminiscent of an inclined ( $\approx70^{\circ}$) disc or ring, although it is difficult to obtain a close fit from the model.

3.14 Arp 220

Again, we experienced difficulties in fitting the model to this spectrum; we could either account for the intensity and the width, but not the shape with a highly inclined disc, or account for the intensity and shape but not the width, with a disc of intermediate inclination. In any case, since a $\sim200''$ beam gives the observed intensity, it is feasible that the molecular ring/disc extends to around 2 kpc. Since the luminosity is slightly greater within the beam than of the ring/disc in Circinus ( 2.6 10³ K km s^-1 kpc², Curran et al. 2000) a slightly larger beam would be required to obtain a match, thus decreasing this distance estimate slightly. From their interferometric observations, Scoville et al. (1997) find that the CO, which is distributed in a thin disc, extends out to $\approx1.4$ kpc. An inclination of $40^{\circ }$ $-50^{\circ}$ is derived for the disc and so we adopt this value.

3.15 NGC 6814

The best match was obtained using a low inclination ( ) ring "observed'' with a 104'' beam, giving a profile similar to that in Fig. 5, although somewhat narrower. This suggests that the extent of the gas does scale as the optical extent in this case, although due to the proximity of this galaxy, this only corresponds to a radius of $\approx1$ kpc. It should be noted that the CO luminosity within the beam is somewhat lower than that in Circinus ( 0.38 10³ K km s^-1 kpc², Young et al. 1995) which would cause a lower observed intensity and so this radius should be regarded as an upper limit only.

3.16 NGC 7130

Since the $1\rightarrow 0$ and $2\rightarrow 1$ observed spectra observed spectra are so narrow (  km s^-1), we expect this disc structure to be of low inclination, and the profile shape is indeed best fitted with a close to face-on disc/ring observed with a

Figure 10: A face-on disc observed with a 160'' beam

160'' beam (as opposed to a $\approx700''$ beam required for a 600 pc ring at 64 Mpc), Fig. 10. This suggests that the ring/disc of molecular gas extends to around 4 times the distance than in Circinus. The luminosity within the beam is slightly greater than that in Circinus ( 2.6 10³ K km s^-1 kpc², Curran et al. 2000) and so a smaller radius may be required in to produce the observed velocity range. This range could feasibly be accounted for with a disc model of higher inclination but the model profile loses its "pointed'' shape.

3.17 NGC 7172

The $2\rightarrow 1$ observed profile has a shape somewhat similar to that of NGC 0034, with the 3-component CO $1\rightarrow 0$ shape of Maiolino et al. (1997) also suggesting an inclined ring, possibly with associated outflow. Testing various models, a $60^{\circ }$ring+outflow model "observed'' with a beam exceeding 200'' gives the best match. Worth noting is that a ring only model (also of $60^{\circ }$ inclination) at a HPBW of 200'' gives the observed intensity, although the central component in the profile is missing. So we summarise the model as being a highly inclined ring in which the CO is confined to  pc with a possible outflow. The presence of the outflow in the model permits an extent of $\approx 600$ pc for the gas. The low intrinsic luminosity ( 0.23 10³ K km s^-1 kpc², Curran et al. 2000) could increase the values of the inferred radii.

3.18 NGC 7469

Although a 2-peaked profile may be feasible in our observed spectrum, and also in that of Heckman et al. (1989), no such feature is apparent in the detections of Maiolino et al. (1997); Papadopoulos & Seaquist (1998). In fact we find that the best fit is provided by a ring/disc model at low inclinations, without the presence of an outflow. Also, our models show that a nearly-face on disc/ring gives the observed profile, but not the required width. A slightly more inclined disc (a ring does not suffice here) gives the required width, although the shape is perhaps only consistent with the detection of Papadopoulos & Seaquist (1998), and the model should be "observed'' with a smaller beam in order to obtain the observed intensity. Since NGC 7469 has a higher CO luminosity in the beam than Circinus ( 2.0 10³ K km s^-1 kpc², Young et al. 1995; Curran et al. 2000), which could widen the profile slightly, we favour the narrower ( $10^{\circ }$ inclination) model. Being a Sy1.2, this would align the ring close to the obscuration and the beam size suggests that the molecular gas extent does (nearly) scale with the optical disc in this case (i.e. about three times more extended than that in Circinus). Heckman et al. (1986); Wilson et al. (1991) derive 1 - few kpc for the radius of the star-burst ring.

3.19 Discussion of the model results

We have simulated the observed profiles using a model based on that of the Circinus ring. It should be noted, however, that the exact profile shapes depend upon the distribution of the gas. For example, because of the flat intensity distribution out to $\approx300$ pc, the model (with no outflow) distribution (Figs. 9 and 10 of Curran et al. 1998) will give a single peaked profile for an inclined disc/ring, Fig. 5. If the beam is significantly larger than this, however, a relative deficit in the low velocity gas and the large negative gradient in the intensity at high velocities (from $\approx300$ pc to $\approx400$ pc) will give a double peaked profile, also shown in Fig. 5. This could feasibly account for the observed profiles for NGCs 0034, 2273 and 7172. Also, although the choice of rotation curve is not crucial to the profile shape (Curran 1998), a constantly rising rotation curve may possibly reproduce the observed spectrum of NGC 5548. Thus we emphasise again that these models can only provide a very rough indicator of the molecular gas distribution. Therefore in the previous discussion (Sects. 3.1-3.18) particular attention should only be paid to the derived inclinations, which define the profile shapes. These are summarised in Table 2, and the values should only be considered as approximate estimates, although where interferometric data is avaliable, i.e. NGC 1068 and Mrk 231, our estimates give comparable results (due to uncertainties in Mrk 273 and Arp 220, we have adopted inclinations from the literature, Table 2).

  

 

Table 2: Summary of the model results. r is the estimated upper limit to the molecular gas extent in relation to Circinus (ring/disc to $\approx 600$ pc). $i_{\rm mol}$ is the inclination of the molecular ring estimated from the models with the Circinus value coming from Curran et al. (1998). $i_{\rm main}$ is the approximate inclination of the main galaxy disc. In most cases these have been calculated (assuming a circular shape) from the ratio of the major and minor optical axes (NASA/IPAC Extragalactic Database); the exceptions are Circinus (Freeman et al. 1977), NGC 1068 (Tacconi et al. 1994) and NGC 1365 (Hjelm & Lindblad 1996), of which only the latter disagrees with the NASA/IPAC estimated value ( $\approx 60^{\circ }$). Again, Sy is the Seyfert type which is expected to be directly related to the inclination of the obscuration; Sy1 face-on, Sy2 edge-on and intermediate classes of intermediate inclinations. In the case of Circinus and NGC 1365 the actual inclinations are given based on ionisation cone/molecular outflow estimates (Curran et al. 1999 and references within; Hjelm & Lindblad 1996, respectively). The value for NGC 1365 is somewhat lower than expected for a Sy1.8, although Hjelm & Lindblad (1996) class this as Sy1.5

Galaxy	r	$i_{\rm mol}$	$i_{\rm main}$	Sy
Circinus	1	$78^{\circ}$	$65^{\circ}$	$78^{\circ}$
NGC 0034	<3		$70^{\circ}$	Sy2
NGC 1068	<2	$40^{\circ}-60^{\circ}$	$35^{\circ}$	Sy2
NGC 1365		$30^{\circ}-40^{\circ}$	$40^{\circ }$	$35^{\circ}$/Sy1.8
NGC 1667		$\sim30^{\circ}$	$40^{\circ }$	Sy2
UGC 03374			$40^{\circ }$	Sy1.5
NGC 2273	$\sim1$	$20^{\circ}-50^{\circ}$	$40^{\circ }$	Sy2
NGC 4593			$40^{\circ }$	Sy1
Mrk 231	<7		$40^{\circ }$	Sy1
NGC 5033	-		$60^{\circ }$	Sy1.9
Mrk 273		$\approx 60^{\circ }$	$75^{\circ}$	Sy2
NGC 5135	<5	$\approx10^{\circ}$	$46^{\circ}$	Sy2
NGC 5347	-	-	$40^{\circ }$	Sy2
NGC 5548	-	$\approx70^{\circ}$	$20^{\circ }$	Sy1.5
Arp 220		$40^{\circ}-50^{\circ}$	$40^{\circ }$	Sy2
NGC 6814	$\sim1$		$20^{\circ }$	Sy1.5
NGC 7130		-	$20^{\circ }$	Sy2
NGC 7172		$\sim60^{\circ}$	$60^{\circ }$	Sy2
NGC 7469	$\sim3$	$\sim10^{\circ}$	$40^{\circ }$	Sy1.2

Addressing the issue of differences in orientation between the galactic disc, large-scale molecular ring and small-scale torus, Circinus, NGCs 0034, 1365, 5033, 7172 and Mrk 273 (which are all Sy2s) all have $i_{\rm main}\approx i_{\rm mol}\approx$ dusty torus inclination (based on Seyfert type), which is consistent with the results of McLeod & Rieke (1995); Wilson & Tsvetanov (1994); Capetti et al. (1996). Examining this in more detail, the upper and lower limits in the inclination angles (Table 2) makes it difficult to select a median value, although for $i_{\rm main}-i_{\rm mol}$ this seems to be around a value of just under $20^{\circ }$, and since many of the inequalities occur at this value, we choose this for the median. For $i_{\rm mol}-{\rm Sy}$, we use the same median and make the following approximations:

1.

For Sy2s, a torus inclination ,

2.

for Sy1s, a torus inclination , and

3.

for Sy1.5s, a torus inclination $\approx45^{\circ}$.

The results are summarised in Fig. 11.

Figure 11: Left: The offset in inclination (degrees) between the main galaxy disc and the molecular disc/ring. Right: The offset in inclination (degrees) between molecular ring and the obscuring torus. Black boxes represent Sy2s and white boxes Sy1s

From Fig. 11 (left), we see that Sy2s tend to have their molecular ring/disc aligned with the main galaxy disc, whereas Sy1s tend to show a misalignment. However, the only Sy1 with an offset is actually a Sy1.5 (NGC 6814) and this value could quite possibly be , Table 2. Note that NGC 7130 (Sy2) is a borderline case with an offset of less than $20^{\circ }$ being quite feasible. Also, possibly consistent with our results, in NGC 1068 Schinnerer et al. (1999) suggest that the molecular disc has a higher inclination than the main disc, Table 2. Concerning the alignment between the dusty torus and the molecular ring, there appears to be no dependence on Seyfert type from our very approximate statistics, although for the sample (except NGCs 1667 and 5135 these seem to be aligned within $\approx30^{\circ}$, Fig. 11 (right).

Our results partly confirm the findings of McLeod & Rieke (1995); Maiolino & Rieke (1995), i.e. the galactic disc is aligned with the molecular ring. As mentioned previously, we find this only to be strictly true for Sy2s, although an offset of $\gg20^{\circ}$ is only seen for one of the whole sample (NGC 5548, in which the model was noted to be uncertain, Sect. 3.13). Consistent with their results, from Table 2 we find that there is a (slight) tendency for Sy1s to be located in galaxies of low inclination and for Sy2s in galaxies of higher inclination, although intermediate inclinations seem to be favoured by both types. Regarding the relative aspects of the ring and torus, perhaps contrary to McLeod & Rieke (1995); Maiolino & Rieke (1995) we find that most tori tend to be aligned with the galactic disc (Table 2), although our result does appear to be consistent with that of Keel (1980); Lawrence & Elvis (1982), and more recently, Wilson & Tsvetanov (1994); Capetti et al. (1996). This suggests, since there are no large ( $\gg20^{\circ}$) misalignments, that the conservation of angular momentum holds each structure approximately coplanar with its neighbour, i.e. and . Concerning particular examples, Gallimore et al. (1999) find from a sample of 13 Seyfert galaxies that the neutral atomic gas is distributed in a 100-pc scale rotating disc which has its axis aligned with that of the host galaxy and Greenhill et al. (1997) find that the masing disc in the Sy2 NGC 4945 has the same position angle as the host disc.