It appears that the unified model needs to be refined in order to better fit the observations. The unified model makes the simplified assumption that all Seyferts have tori, and that all tori are exactly alike, with the same relative alignment to the inner accretion disk and the outer galactic disk. The true nature of the obscuring medium is little understood. What are the dimensions of the tori? How large are the opening angles? Do they all have the same column densities? Are they present in all Seyfert galaxies? Are they always the source of the attenuation? Are they coplanar with the accretion disks and the galactic spirals?
Schmitt et al. [Schmitt et al., 1997] have probed this last question. They propose that whether we see a type 1 or a type 2 Seyfert galaxy does depend upon our viewing angle, but not solely upon whether our line of sight towards the central engine runs through the putative torus. Instead, they suggest that the torus' axis may not be coupled with that of the host galaxy, and a type 2 is observed when our line of sight runs through a large column of matter either in the torus or in the disk of the host galaxy itself. In other words, the necessary condition for observing a Seyfert 1 object is that the axes of the torus and the galaxy must lie close enough to both each other and to our line of sight in order to provide us with an unimpeded view of the innermost regions; if the axes are too far apart, then the direct emission of the central region will be absorbed in most directions. Supporting their hypothesis, they find that 14 of 15 Seyfert 1's have torus and galaxy axes no more than 60 degrees apart, while the axes in 31 Seyfert 2's have an almost random relative orientation.
There are some observations which lend support to the Schmitt et al.
hypothesis. More than ten years ago, Ulvestad and Wilson found
that the tori axes (assumed to be parallel to the radio elongation)
and the radio axes in a distance limited sample have random relative
orientation [Ulvestad and Wilson, 1984]. A few years later, Schulz found an
extended NLR in NGC 4151 whose velocity matches the galactic rotation
curve, suggesting that the extended NLR is made of galactic disk
material and therefore the direction of the nuclear collimation is
into the plane of the galaxy [Schulz, 1988]. More recently,
Reynolds and collaborators found that the X-ray spectrum of the
Seyfert 2 galaxy NGC 6552 is best explained by a dominant contribution
from reflection off of a cold, neutral gas seen nearly face on. This
could be the cold outer parts of an accretion disk seen face on, which
would mesh nicely with the recent discovery of emission from the hot
inner regions of face on accretion disks within Seyfert 2's by Turner
et al. Alternately they suggest that a bar-like structure seen in the
galactic disk may be due to the collimated light of the nucleus, and
therefore suggest that the cold reflection region may be material in
the plane of the galaxy rather than a molecular torus [Reynolds et al., 1994].
Also, X-ray spectrum of the Circinus galaxy has been found to be
dominated by a single neutral reflector; unlike NGC 1068, there does
not appear to be a significant Fe K- 
feature due to the warm
scattering region[Matt et al., 1996]. This lack of an observed electron
scattering component suggests that the cold absorber blankets the
entire nuclear region, and therefore may be due to the disk of the
host galaxy.