Äîêóìåíò âçÿò èç êýøà ïîèñêîâîé ìàøèíû. Àäðåñ îðèãèíàëüíîãî äîêóìåíòà : http://sn.sai.msu.ru/~sil/preprints/n7217_aa.ps.gz Äàòà èçìåíåíèÿ: Fri Feb 2 00:25:00 2001 Äàòà èíäåêñèðîâàíèÿ: Mon Oct 1 19:45:45 2012 Êîäèðîâêà: Ïîèñêîâûå ñëîâà: ï ï ï ï ï ï ï ï ï ð ï ð ï ð ï ð ï ð ï ð ï

Astron. Astrophys. 364, 479--490 (2000) ASTRONOMY
AND
ASTROPHYSICS
Decoupled nuclei and nuclear polar rings in regular spiral galaxies
NGC 7217 #
O.K. Sil'chenko 1,##,### and V.L. Afanasiev 2,+
1 Sternberg Astronomical Institute, University av. 13, Moscow 119899, Russia
2 Special Astrophysical Observatory, Nizhnij Arkhyz, 357147 Russia
Received 21 July 2000 / Accepted 26 September 2000
Abstract. The regular isolated Sab galaxy NGC 7217 has been
studied with the Multi­Pupil Fiber Spectrograph of the 6m tele­
scope of the Special Astrophysical Observatory RAS (Nizhnij
Arkhyz, Russia) in two spectral ranges, the blue one including
the strong absorption lines Mg I and Fe I and the red one in­
cluding the emission lines H# and [N II]#6583. We confirm
the existence of a circumnuclear gaseous polar disk with a ra­
dius of 3 ## which we reported earlier. The same area, with a
radius of 3 ## --4 ## , elongated orthogonally to the line of nodes, is
distinguished by high values of the Lick index #Fe# and shows a
Mg/Fe ratio lower than solar. This implies that there were at least
two discrete star formation bursts in the circumnuclear region
with a temporal separation of a few Gyrs. We relate this pair of
bursts to the complex structure of the global brightness profile
of the galaxy, which may be decomposed into three exponential
segments with different scalelengths.
Key words: galaxies: kinematics and dynamics -- galaxies:
structure -- galaxies: spiral -- galaxies: evolution -- galaxies: nu­
clei -- galaxies: photometry -- galaxies: individual: NGC 7217
1. Introduction
A careful investigation of the inner parts in normal spiral galax­
ies with modern techniques reveals almost always a compli­
cated structure which cannot be described by a classical two­
component model consisting of a centrally concentrated, old,
red, smooth de Vaucouleurs' bulge and a more extended, young,
blue exponential disk with spiral arms. Among recent find­
ings which may have serious implications for a global evolu­
tionary picture we can mention papers on exponential bulges
Send offprint requests to: O.K. Sil'chenko (olga@sai.msu.su)
# Partly based on observations collected with the 6m telescope of the
Special Astrophysical Observatory (SAO) of the Russian Academy
of Sciences (RAS) which is operated under the financial support of
Science Department of Russia (registration number 01­43) and on data
from the ING Archive and the HST Archive.
## Isaac Newton Institute of Chile, Moscow Branch
### UK Astronomy Data Centre, Guest Investigator
+ Isaac Newton Institute of Chile, SAO Branch
Table 1. Global parameters of NGC 7217
Hubble type (R)SA(r)ab
R25 9.3 kpc
B 0
T 10.53
MB --20.5
(B - V ) 0
T 0.77
(U -B) 0
T 0.25
Vr 952 km · s -1
Distance 16.4 Mpc (H0=75 km · s -1 · Mpc -1 )
i(phot) 35 o
PA(phot) 95 o
(Andredakis et al. 1995, Courteau et al. 1996), multi­tier disks
-- e.g. in NGC 5533 (Sil'chenko et al. 1997a) or in NGC 157
(Ryder et al. 1998), decoupled circumnuclear stellar disks (e.g.
in NGC 4594, Emsellem et al. 1996), inner polar gaseous disks
(e.g. in NGC 2841, Sil'chenko et al. 1997b), circumnuclear spi­
ral arms (e.g. in NGC 5248, Laine et al. 1999, or in NGC 488,
Sil'chenko 1999), etc. In some early­type spiral galaxies, e.g.
in NGC 4138 (Jore et al. 1996) and in NGC 7217 (Merrifield
& Kuijken 1994), the existence of counterrotating stars in their
global disks was claimed. This may mean that discrete catas­
trophic events, such as minor mergers or gas accretion from a
satellite, govern the evolution of even quite regular galaxies.
But there may also exist some intrinsic evolutionary processes,
such as bar formation and dissolution due to disk instabilities or
mass concentration in the center, that can provide the same set
of properties without any external action. We cannot yet prove
either of these hypotheses because of the lack of detailed inves­
tigations of even nearby galaxies. So any new, careful analysis
of the dynamics, structure, and stellar populations in normal
spiral galaxies gives an additional chance to understand their
evolution.
Here, we present a detailed study of the regular isolated
Sab galaxy NGC 7217, the global parameters of which are
given in Table 1. The galaxy has a prominent bulge and a
truly flocculent spiral structure: unlike some other flocculent
galaxies, e.g. NGC 2841 (Block et al. 1996) or NGC 5055
(Thornley & Mundy 1997) which unveil grand design in the
NIR K­band, NGC 7217 preserves its non­wave appearance

480 O.K. Sil'chenko & V.L. Afanasiev: Decoupled nuclei and nuclear polar rings in regular spiral galaxies
Table 2. Spectral observations of NGC 7217
Date Telescope Configuration Exposure Scale Spectral range Dispersion PA(top)
19.08.98 6m BTA MPFS+CCD 1024 â 1024 60 min 1. ## 0 per lens 4250--5600 š
A 1.3 š
A /px 215 #
19.08.98 6m BTA MPFS+CCD 1024 â 1024 60 min 1. ## 0 per lens 5650--7000 š
A 1.3 š
A /px 215 #
06.01.91 4.2m WHT ISIS+CCD 800 â 1180 25 min 0. ## 335 per px 8000--8840 š
A 0.74 š
A /px 240 #
07.01.91 4.2m WHT ISIS+CCD 800 â 1180 20 min 0. ## 335 per px 5830--6680 š
A 0.74 š
A /px 150 #
07.01.91 4.2m WHT ISIS+CCD 800 â 1180 20 min 0. ## 335 per px 5830--6680 š
A 0.74 š
A /px 240 #
06.11.97 4.2m WHT ISIS+CCD 1024 â 1024 33 min 0. ## 335 per px 6150--7000 š
A 0.8 š
A /px 68 #
even at 2 µm (Elmegreen et al. 1999). But it has another puz­
zling morphological feature, namely, three stellar rings at radii
of 12 ## , 32 ## , and 77 ## , the innermost of which is accompanied by
H# emission enhancement and the outermost also by HI con­
centration, so both are sites of intense star formation (Buta et
al. 1995). Such star forming rings are usually treated as reso­
nance loci of a bar. But morphologically NGC 7217 is a purely
unbarred galaxy; besides, a bar should produce a wave spiral pat­
tern which is not observed in this galaxy. Buta et al. (1995) tried
to solve this problem, through a Fourier analysis of the I­band
image of NGC 7217, and found a modem=2; the conclusion was
that the galaxy has a low­contrast bar aligned along PA # 60 # .
Perhaps, this bar is observed in the late stage of dissolution:
as Athanassoula (1996) has discussed, the rings are more long­
lived than the bars, which may be destroyed by a central mass
concentration. But the most striking peculiarity has been found
by Merrifield &Kuijken (1994): they have claimed the existence
of a counterrotating stellar disk. More exactly, from the anal­
ysis of stellar line­of­sight velocity distribution (LOSVD) data
they noted a counterrotating ``wing'' over the full radius range,
which they have studied, namely from R = 10 ## to R = 60 ## .
This ``wing'' contains 30% of all stars and does not become
weaker with radius; so they conclude that it is not a bulge, but
a counterrotating fraction of the global disk. Immediately, the
question has been raised about secondary gas infall or a minor
merger, though the galaxy looks quite isolated.
We have already studied this interesting galaxy (Zasov &
Sil'chenko 1997): from the ionized gas kinematics in the cir­
cumnuclear region and from an analysis of the central isophotes
we have detected an inclined gaseous disk in the very center of
NGC 7217 and reported a mass concentration of 6 · 10 7 M
# in
its nucleus. During the last years our understanding was grow­
ing on how the phenomena of chemically distinct galactic cores
(Sil'chenko et al. 1992) and kinematically decoupled gaseous
or stellar subsystems in disk galaxies may be related, both be­
ing produced by a secondary gas accretion or minor merger
event. For example, in NGC 2841 we have found a chemi­
cally distinct nucleus, a circumnuclear polar disk of ionized gas
with a radius of 200 pc, and counterrotating stars in the bulge
(Sil'chenko et al. 1997b, Afanasiev & Sil'chenko 1999). The
similarity between the gas and stellar kinematics in NGC 2841
and NGC 7217 has stimulated us to search for a chemically
distinct nucleus in the latter galaxy. The results of this search,
together with additional kinematical and photometric analyses,
are presented in this paper.
2. Observations and data reduction
In 1998 we have undertaken two­dimensional spectroscopy of
NGC 7217 with the Multi­Pupil Fiber Spectrograph (MPFS)
of the 6m telescope of the Special Astrophysical Observatory
(Nizhnij Arkhyz, Russia). Two spectral ranges were exposed:
a blue­green one, 4250--5600 š
A, and a red one, 5650--7000 š
A.
The detailed parameters of the spectral observations are given
in Table 2. A grating of 1200 grooves per mm was used which
provided a reciprocal dispersion of 1.3 š
A per pixel and a spectral
resolution of 3 š
A. The seeing of FWHM=1. ## 6 was estimated
from a stellar exposure.
These spectral observations have been made with the new
variant of the panoramic spectrophotometer which became
operational at the prime focus of the 6m telescope in the
end of 1997. With respect to the previous variants of MPFS
(Afanasiev et al. 1990, Afanasiev et al. 1996), the field of view
is now increased and the common spectral range is larger due
to the use of fibers: they transmit light from 16 â 16 square
elements of the galaxy image to the slit of the spectrograph
(256 fibers) together with the sky background taken 4. # 5 away
from the galaxy itself (6 fibers). The size of one spatial ele­
ment is 1 ## â 1 ## . At the exit of the spectrograph a 1024 â 1024
CCD registers all 262 spectra simultaneously. The primary re­
duction of the data is made within IDL. After bias subtracting,
flatfielding, and one­dimensional spectra extraction from the
CCD frame, we linearize and analyse each spectrum individ­
ually. The one­element spectral characteristics, such as fluxes
in continuum or in emission lines, redshift, and absorption­line
indices are then combined into two­dimensional arrays corre­
sponding to the galactic region under consideration with the
help of software developed earlier in the Special Astrophysi­
cal Observatory (Vlasyuk 1993) and with our own programs.
To calculate absorption­line indices and their errors we have
used also the program of Dr. Vazdekis. As a result, we ob­
tain two­dimensional surface brightness distributions, veloc­
ity fields, and maps of stellar population characteristics. In the
blue­green spectral range, we measure the absorption­line in­
dices H#, Mgb, Mg 2 , Fe5270, and Fe5335 in the popular Lick
system (Worthey et al. 1994); to check the consistency of our
measurements with the model indices calculated in this sys­
tem (Worthey 1994), we also observed stars from their list
(Worthey et al. 1994). Besides that, we use our blue­green spec­
tra to derive a stellar velocity field in the center of NGC 7217
by cross­correlating elementary galactic spectra with the spec­
trum of a K­giant star -- the brighter component of the visual

O.K. Sil'chenko & V.L. Afanasiev: Decoupled nuclei and nuclear polar rings in regular spiral galaxies 481
binary STF 2788. In the red spectral range we have measured
baricentric positions of the emission line [N II]#6583, which is
the strongest in the center of NGC 7217, to derive a velocity
field of the ionized gas. We have estimated the best accuracy of
our velocity measurements as 10 km s -1 from the night­sky line
[OI]#6300 analysis. For the absorption­line index accuracy, we
have made estimates using the method of Cardiel et al. (1998):
the typical error of the indices varies for the EW­like indices
from 0.15 š
A in the nucleus to 0.5 š
A in the individual elements
at the edges of the area investigated, and from 0.004 to 0.01 for
Mg 2 . To keep a constant level of accuracy along the radius, we
summed the spectra in concentric rings centered on the nucleus
and studied the radial dependencies of the absorption­line in­
dices by comparing them to the synthetic models of old stellar
populations of Worthey (1994) and Tantalo et al. (1998). We es­
timate the mean accuracy of our azimuthally­averaged indices
as 0.1 š
A.
Also, we have taken several long­slit spectra of NGC 7217,
which have been obtained at the William Herschel Telescope on
LaPalma with the ISIS, from the ING Archive. The details of ex­
posures are also given in Table 2. These CCD frames have been
reduced with the software of Dr. Valeri Vlasyuk (Vlasyuk 1993).
The photometric data involved in our analysis are taken
from the ING and HST Archives. The broad­band I image of
NGC 7217 has been obtained on June 1st, 1998, at the 1m Ja­
cobus Kapteyn Telescope onLaPalma.The exposure times were
10 min, 10 min, and 5 min, but only the first of the exposures
was well guided; only this is analysed in this work. The seeing
quality is estimated from neighbouring star measurements as
FWHM # = 1. ## 5. The central part of the galaxy has been also
observed by the Hubble Space Telescope. The earlier observa­
tions with WFPC2 were made on June 10, 1994, through the
filter F547M, with an exposure time of 5 min (Principal Inves­
tigator: W. Sargent, Program ID: 5419). Later, it was observed
with the NICMOS2 through the filters F110W and F160W dur­
ing 128 sec each on August 17, 1997 (Principal Investigator: M.
Stiavelli, Program ID: 7331). The spatial resolution was 0. ## 1 for
WFPC2 observations and 0. ## 2 for the NICMOS observations.
We have derived morphological characteristics of the surface
brightness distribution in NGC 7217 by analysing these images.
The program FITELL of Dr. Vlasyuk has been used for tracing
the isophote major axis position angle and ellipticity along the
radius, and 2D image decomposition was performed with the
software FVIZ and IMAR (Vlasyuk 1993) as well as with our
own programs.
3. A magnesium distinct nucleus
and a Fe­rich circumnuclear disk
To study properties of the stellar populations in the central re­
gions of galaxies, we use metal­ and hydrogen­line indices con­
fined to a rather narrow green spectral range, namely, H#, Mgb,
Fe5270, and Fe5335. In the case of NGC 7217 there are some
problems with these indices, however. First of all, rather intense
H# emission is observed in the center of the galaxy, except in
the narrow radial range of R = 4 ## - 6 ## . Therefore one can ex­
pect that the absorption­line index H#, which is a good indicator
of stellar population age, will be strongly contaminated by the
emission almost everywhere and therefore cannot be used. Sec­
ondly, in the nucleus itself a noticeable emission line [N I]#5200
is seen, and as Goudfrooij & Emsellem (1996) noted, this emis­
sion causes an overestimation of the absorption­line index Mgb
because it falls into the continuum band of this index. There­
fore we are forced to use an index Mg 2 which has a broader
continuum base than Mgb, but is more dependent on the correct
calibration of the global spectrum shape. To calculate Mg 2 , we
use galactic spectra calibrated into absolute fluxes.
Fig. 1 shows the isocontours of two­dimensional distribu­
tions of the Lick absorption­line indices, Mg 2 (left) and #Fe# #
(Fe5270+Fe5335)/2 (right) in the central 16 ## of NGC 7217.
The magnesium index is strongly peaked near the nucleus; it de­
creases rather symmetrically with radius, and the shape of outer
Mg 2 contours is not too different from the shape of continuum
isophotes. However, in the very center an elongated structure
seems to appear at the limit of the spatial resolution; it is a bit
shifted to the east from the center of the brightness distribution.
The distribution of the iron index is much more complicated. In
the patchy pattern, one can distinguish an elongated structure of
enhanced #Fe# # (Fe5270+Fe5335)/2, also shifted to the east
from the nucleus and much more extended than the analogous
structure in the Mg 2 map: it can be traced up to R # 5 ## . We
shall refer to it as the `Fe­rich disk'; in the next sections we will
argue that it can be indeed related to the circumnuclear stellar
disk. However, let us note here that the structure with the en­
hanced metal absorption lines is elongated in the south­north
direction whereas the global line of nodes is oriented approxi­
mately east­west (see Table 1).
Lick absorption­line indices have been well calibrated with
respect to the integrated (luminosity­averaged) properties of
stellar populations in numerous works using evolutionary syn­
thesis techniques. Our conclusions derived below are based on
the models of Worthey (1994) and Tantalo et al. (1998) for
single­age single­metallicity stellar populations. Diagnostic di­
agrams of index vs. index are presented in Fig. 2. The diagram
#Fe# vs. Mg 2 (left) is being considered by many specialists in
chemical evolution as a tool to limit the duration of the main star
formation episode (see, e.g., Matteucci 1994). WhenWorthey et
al. (1992) noted for the first time a magnesium overabundance
in most bright elliptical galaxies, this phenomenon was soon
treated as a natural consequence of the short duration of their
star forming epoch, less than 1 Gyr. The nature of this effect is
deduced from the theoretical prediction that magnesium has to
be produced mainly by SNeII, which explode earlier than the
bulk of the iron­producing SNeIa from the same stellar gen­
eration; if the star formation process stops between these two
moments, the stars would have a higher magnesium­to­iron ra­
tio than the solar, which corresponds to continuous star forma­
tion. The star formation histories in the centers of early­type
disk galaxies are less clear than those in ellipticals, and ob­
servational results on their Mg/Fe ratios are also contradictory.
Our statistics (Sil'chenko 1993) showed evidence for a solar
Mg/Fe ratio in the centers of almost all disk galaxies, from Sc

482 O.K. Sil'chenko & V.L. Afanasiev: Decoupled nuclei and nuclear polar rings in regular spiral galaxies
Fig. 1. Two­dimensional maps of the Lick absorption­line indices (isocontours) in the central region of NGC 7217: Mg 2 (left, the outermost
isocontour corresponds to 0.180, the step between isocontours is 0.005) and #Fe# (right, the outermost isocontour corresponds to 2.50 š
A, the
step between isocontours is 0.1 š
A). The gray­scaled background is the distribution of the green continuum.
Fig. 2. Comparison of our observational data for NGC 7217 with the models of Worthey (1994) and Tantalo et al. (1998) for [Mg/Fe]=0. The
ages of the models in the legend are given in billion years. The observational points connected by a dashed line -- bells in the diagram #Fe#
vs Mg 2 (left panel) and circles in the diagram H# vs #Fe# (right panel) -- are azimuthally averages and taken along the radius with a step of
1. ## 0, the bulge points starting from R = 6 ## ; the points for the Fe­rich disk in the left panel are connected from its northern end to the southern
one; unconnected observational points are for individual elements. The metallicities for Worthey's models are +0.50, +0.25, 0.00, --0.22, --0.50,
--1.00, if one takes the signs along the sequences from the right to the left, and for the models of Tantalo et al. they are +0.4, 0.0, and ­0.7. The
mean data for ellipticals of intermediate luminosity on the left panel are taken from Trager et al. (1998).
to S0, in both bulge­ and disk­dominated objects. Jablonka et
al. (1996) reported a strong magnesium overabundance, up to
[Mg/Fe]# +0.5, in bulges brighter than M r # -19.5, includ­
ing our target galaxy NGC 7217. Our data in Fig. 2 (left) differs
from those of Jablonka et al.: here, the nucleus of NGC 7217 lies
near the model sequences of Worthey (1994) implying a solar
Mg/Fe ratio or even a slight iron overabundance. This result is
due to a smaller Mg 2 estimate, 0.25 instead of 0.28 in Jablonka
et al. (1996), but mainly to a higher Fe5335 value, 3.0 š
A instead
of 2.4 š
A. Curiously, the estimate of Fe5270 made by Jablonka
et al. (1996) is quite consistent with ours. We cannot explain
the cause of the agreement in Fe5270 and the disagreement in
Fe5335 but wewould like to note that in model calculations (e.g.
see Worthey 1994) the values of Fe5270 and Fe5335 obtained
are almost equal to each other, as is the case in our data.

O.K. Sil'chenko & V.L. Afanasiev: Decoupled nuclei and nuclear polar rings in regular spiral galaxies 483
The points related to the bulge of NGC 7217 are dispersed
around the model sequences with a solar magnesium--to--iron
ratio, and none of them fall into the area occupied by el­
lipticals of similar luminosity (MB = -19 Â -20), whose
mean locus is plotted using the data from Trager et al. (1998)
(a shaded horizontal band in Fig. 2, left). The azimuthally­
averaged measurements for the bulge also lie along the se­
quences for [Mg/Fe]=0. But the most interesting behaviour is
demonstrated by the points taken within the Fe­rich circum­
nuclear disk. In this rather bright, central part of the galaxy,
the accuracy of the arcsecond element indices, of, say, #Fe#, is
better than 0.3 š
A. The deviations of the extreme points of the
Fe­rich disk at R # 5 ## from the model sequence [Mg/Fe]=0
are at the 3# level, and certainly due to iron overabundance.
This is a rather puzzling observational phenomenon common
for some types of irregular galaxies (e.g. the LMC), and it is
usually treated as evidence for a bursty character of the star
formation (Gilmore & Wyse 1991, Marconi et al. 1994). When
approaching the center of the disk, the Mg/Fe ratio rises and
near the nucleus it becomes close to solar. The total metallicity
of the circumnuclear disk is higher than that of the bulge. Ob­
viously, the Fe­rich disk is a secondary formation product. The
star formation within it had to be strongly inhomogeneous, with
an exotic mechanism to result in iron overabundance at its outer
edge: a kind of transient star forming circumnuclear ring.
As we mentioned earlier, the age diagnostics in the center
of NGC 7217 is complicated by the noticeable Balmer emis­
sion contaminating the Lick index H#. So in the diagram H#
vs. #Fe# (Fig. 2, right) the majority of the observational points
trace only upper limits for the age of the stellar population.
But in the narrow gap between the LINER nucleus and the star
forming ring (see Sect. 5), at R = 4 ## - 6 ## the H# emission is
absent and we can use the H# absorption­line index to estimate
a luminosity­weighted age of the bulge stellar population. From
Fig. 2 (right) one can see that the bulge of NGC 7217 is rather
old, at least 10 Gyr. The ages of the nucleus and of the Fe­rich
circumnuclear stellar disk cannot be determined exactly.
4. Kinematics of stars and gas in NGC 7217
4.1. Circumnuclear region
In the previous paper (Zasov & Sil'chenko 1997) we noted
clear signatures of the dynamical distinctness of the nucleus in
NGC 7217: the rotation axis of the ionized gas inside R # 5 ##
turns by almost 90 # . Now we present further proofs of the ex­
istence of a circumnuclear `polar' gaseous ring (disk?) in this
galaxy.
We now proceed to determine the parameters of the stellar
and gaseous rotation around the nucleus of NGC 7217. Under
the assumption of planar axisymmetric rotation, the azimuthal
dependence of central line­of­sight velocity gradients is:
dv r /dr = # sin i cos(PA - PA 0 ),
where # is the deprojected central angular rotation velocity, i is
the inclination of the rotation plane and PA 0 is the orientation
of the line of nodes, coinciding in the case of an axisymmetric
ellipsoid (or a thin disk) with the photometric major axis. So by
fitting azimuthal variations of the central line­of­sight velocity
gradients with a cosine curve, we can determine the orientation
of the dynamical major axis by its phase and the central angular
rotation velocity by its amplitude. To apply this procedure, we
need two­dimensional velocity fields.
The new MPFS observations allow us to construct two­
dimensional velocity maps for the central 16 ## â 16 ## region
of NGC 7217. Fig. 3 shows such maps for the stars (left) and for
the ionized gas as traced by the [N II]#6583 emission (middle),
as well as the stellar velocity dispersion map (right). One can
see that two velocity maps look different. At first glance, the dy­
namical major axis of the stellar component seems to be close to
the global line of nodes, west­east. However it is strange that the
brightness center coincides with a flat area in the velocity map.
If the dynamical center coincides with the brightness center,
we would obtain a central projected angular rotation velocity
of only (8 ± 2) km/s/arcsec. If we shift the dynamical center 3 ##
to the east, to the point of maximum isovelocity crowding, we
can increase # proj # # sin(i) to (22 ± 5) km/s/arcsec. Fortu­
nately, the direction of the maximum line­of­sight velocity gra­
dient, or the dynamical major axis, remains the same for both
dynamical center locations: it is PA dyn = 241 # ± 2 # within
R = 3 ## , some 30 # different from the line­of­nodes direction,
but coinciding with the PA of the low­contrast bar reported by
Buta et al. (1995). Outside R # 3 ## the dynamical major axis
turns to PA dyn = 276 # , implying an axisymmetric stellar rota­
tion parallel to the symmetry plane of the galaxy. The velocity
map for the ionized gas obtained by measuring the baricenters
of the [N II] emission line (H# is less relevant because of the
underlying absorption line) looks even more striking than that
for the stars. The isovelocities in the very center demonstrate
a strong so­called ``S­shape'' distortion, whose dynamical ma­
jor axis direction changes by 90 # in the radius range of 3 ## --7 ## .
By fitting a cosine curve to the line­of­sight velocity gradients
within 3 ## from the center, we obtain a projected angular rota­
tion velocity of (17±7) km/s/arcsec and a dynamical major axis
PA dyn = -31 # ± 4 # , orthogonal to the dynamical major axis
of the stars in this radius range. Again, we find the signature of
a circumnuclear polar gaseous disk.
The two­dimensional map of the stellar velocity dispersion
in the central 16 ## region (Fig. 3, right) looks somewhat peculiar:
though rather smooth, it demonstrates an elongated `saddle' of
relatively low # # , at about # # # 130 km s -1 , in PA # 60 # and
a slight increase of # # along the kinematic minor axis on both
sides of the nucleus. We have never seen anything like this and
cannot give an interpretation of this map. The dynamical center
which we define here as a center of symmetry of the stellar
velocity dispersion map again seems to be shifted to the east of
the brightness center.
We can check our result on the decoupled rotation of the
circumnuclear ionized gas with the long­slit spectral data from
the ING Archive. Fig. 4 shows line­of­sight velocities measured
from [N II]#6583 and H# emission lines in PA = 68 # and
PA = 150 # . In PA = 68 # , not too far from the line of nodes,

484 O.K. Sil'chenko & V.L. Afanasiev: Decoupled nuclei and nuclear polar rings in regular spiral galaxies
Fig. 3. Two­dimensional line­of­sight velocity fields for the stars (left) and for the ionized gas (middle), and the map of the stellar velocity
dispersion (right) in the central 16 ## â 16 ## of NGC 7217. The gray­scaled background represents the continuum distribution in the green for
the stars and in the red for the gas.
Fig. 4. Long­slit position­velocity cross­sections in PA = 68 # , not too far from the line of nodes, (left) and PA = 150 # (right) for the center
of NGC 7217; [N II] and H# emission line measurements are plotted.
one can see a flat velocity profile segment in the very center,
till ±3 ## from the nucleus, implying that the dynamical major
axis of the circumnuclear ionized gas is aligned orthogonally, in
PA # 158 # (or --22 # ). Outside this central region the gas shows
much larger projected velocities with respect to the nucleus. In
PA = 150 # , consistent with the previous cross­section, the
central part of the velocity profile shows fast decoupled rotation
within ±4 ## from the nucleus with a slope of 24 km/s/arcsec --
evidently, this direction is close to the dynamical major axis of
the circumnuclear gas; outside the radius of 4 ## the projected
rotation velocity falls to a half of the maximum value reached at
the 4 ## radius. Therefore, the long­slit cross­sections confirm the
dynamical distinctness of the central region of NGC 7217 within
a radius of 3 ## --4 ## and the existence of a `polar' circumnuclear
gaseous disk rotating in a plane orthogonal to the global plane
of the galaxy.
4.2. The whole galaxy
Kinematics of the ionized gas in NGC 7217 has been studied
more than once. Long­slit cross­sections along the major axis
were obtained by Peterson et al. (1978), Rubin et al. (1985),
and Buta et al. (1995), and our team also observed NGC 7217
with a long­slit spectrograph and a scanning Fabry­Perot inter­
ferometer at the 6m telescope (Zasov & Sil'chenko 1997). We
have now analyzed the H# and [N II]#6583 emission lines in
the three additional long­slit cross­sections taken from the ING
Archive. The results obtained over the full radius range are quite

O.K. Sil'chenko & V.L. Afanasiev: Decoupled nuclei and nuclear polar rings in regular spiral galaxies 485
Fig. 5. Long­slit cross­sections in PA = 240 # (left) and PA = 150 # (right) of NGC 7217; stellar absorption line measurements obtained by
cross­correlation with the template star spectrum are plotted.
Fig. 6. Long­slit cross­sections in PA = 240 # (left) and PA = 150 # (right) of NGC 7217; stellar velocity dispersions obtained by cross­
correlation with the template star spectrum are plotted. Also the cuts simulated with the MPFS data in the corresponing position angles are
given.
consistent with the velocity curves published previously, and we
do not present the raw velocity data here.
The stellar kinematics in NGC 7217 have received much
less attention than the gas kinematics. Merrifield & Kuijken
(1994) cross­correlated absorption spectra taken along the ma­
jor and minor axes in the spectral range near Mgb#5175 with
those of template stars, and reported a clear evidence of the
presence of a counterrotating stellar subsystem in the disk of
NGC 7217: LOSVDs of stars look double­peaked in the full
radius range considered by Merrifield & Kuijken (1994), i.e.,
up to 60 ## from the center. However, their work lacks direct
measurements of the velocities of the co­rotating and counter­
rotating components, after their separation in the spectra, as well
as estimates of velocity dispersions, so their conclusion that the
counter­rotating component is a part of the global disk of the
galaxy is not very convincing, as it is only based on an am­
biguous photometric decomposition. We have calculated stellar
velocity profiles for two spectra in PA = 240 # , one near the
Na I#5890,5896 lines and another near the Ca II Ir triplet, and
for one spectrum near Na I in PA = 150 # through an interac­
tive Gauss analysis of the cross­correlation peaks (as a template
star, we have taken HR 661). The results are presented in Figs. 5
and 6. We confirm the existence of the counter­rotating stellar
component seen in the cross­sections not far from the major
axis, in PA = 240 # . But its velocity profile (Fig. 5, left) does
not represent a simple mirror picture of the main co­rotating
component. While the main rotation curve is rather flat between
R # 10 ## and 30 ## , the counter­rotating component has a peak
near R = 10 ## - 15 ## and then already falls back to the sys­
temic velocity at R # 20 ## . If the result obtained by Merrifield
& Kuijken (1994) for major­axis cross­section differing by 30 #
in PA from ours is true, i.e., if two maxima of LOSVD are lo­
cated symmetrically with respect to the systemic velocity, this
may signify that the planes of rotation of the co­rotating and
counter­rotating stars are different. The counter­rotating stellar
subsystem may rotate faster than the co­rotating one; the rea­
son for this can be found in Fig. 6. The stellar velocity disper­
sion in NGC 7217 is strongly peaked near the nucleus where it
reaches about 160 km s -1 (previous estimates: 185±18km s -1 ,
Whitemore et al. 1979; 122±3km s -1 , Dressler 1984). In both

486 O.K. Sil'chenko & V.L. Afanasiev: Decoupled nuclei and nuclear polar rings in regular spiral galaxies
Fig. 7. Deprojected rotation curves for the ionized gas (lines) and for
the stars (signes) in NGC 7217. The east and west halves of the galaxy
are shown by different signes
directions, in PA = 240 # and PA = 150 # , it falls rapidly with
radius and becomes less than 100 km s -1 at R # 5 ## . But the
velocity dispersion of the counter­rotating component seems to
be systematically lower than that of the co­rotating one, so since
both rotate in the same potential, the former has the faster ro­
tation. As the stellar velocity dispersion of the counter­rotating
component is of order of 50 km s -1 , we can surely state that it is
in a disk, so the guess of Merrifield & Kuijken (1994) has been
correct.
In Fig. 7 we compare the stellar and gaseous rotation curves
deprojected with the orientation parameters PA 0 = 85 # and
i = 35 # from the data noted in the legend. All three gas ro­
tation curves, the one from Peterson et al. (1978) as approxi­
mated by the analytic formula in the paper of Merrifield and
Kuijken (1994) and our two, from the WHT­91 cross­section in
PA = 240 # and from the WHT­97 cross­section in PA = 68 # ,
are in general agreement with each other: they are flat, with a
maximum velocity level of v rot # 270 km s -1 and a possible
decrease by some 20 km s -1 in the radius range of 4.5--6 kpc.
The stellar rotation looks more irregular though we have taken
into consideration the main component only, i.e. the brighter
which is co­rotating with the gas. The highest values of stellar
rotation velocities are comparable to the gas ones, thus sup­
porting the idea of disk predominance at already R # 1 kpc --
in favour of Kent's (1986) photometric decomposition (see the
next section). But in many places the stellar rotation velocities
fall below the gas rotation curve. The most striking feature of
this kind is seen around R = 3 kpc: at different directions and
wavelengths (implying: for stars of different ages?) the stellar
velocity is 50% --100% less. Interestingly, a similar detail at the
same radius was found by us in the azimuthally averaged rota­
tion curve obtained for the ionized gas with the scanning Fabry­
Perot interferometer (Zasov & Sil'chenko 1997). We noted that
this feature is located at the outer edge of the inner pseudoring
and perhaps related to some non­circular motions; but no de­
tailed interpretation was given. Now, when we detect an even
stronger response to the ring in a dissipationless dynamical sub­
system, we are even more astonished about the nature of the
inner pseudoring.
5. Surface photometry and central structure of NGC 7217
5.1. The central region of NGC 7217
From our red exposure with the MPFS we have derived the
surface distributions of the emission­line intensities of H# and
[N II]#6583, as well as of their ratio (see Fig. 8). NGC 7217 is a
LINER, so the nitrogen emission is the strongest in its center; its
distribution (Fig. 8, top) is concentrated towards the center, but
still well resolved. The inner isophotes of the nitrogen emission
seem to be elongated approximately north­south, i.e. orthogo­
nally to the global line of nodes, but close to the dynamical major
axis orientation found in the previous section for the circumnu­
clear ionized gas. The surface distribution of the H# emission
(Fig. 8, middle) differs from that of [N II]#6583: it demonstrates
an emission ring with the radius of 11 ## , as previously found by
Pogge (1989). Besides the ring, a central concentration also
aligned along the north­south direction is present. The [N II]--
to--H# ratio (Fig. 8, bottom) varies significantly over a tiny area
with a radius of 11 ## : it is 0.5 in the H#­ring, slightly above 1
at the inner edge of the H#­ring, 4 in the nucleus and 5--10 at
the edge of the central emission concentration. We can easily
understand the [N II]--to--H# ratios in the ring and in the nu­
cleus, as they are typical for a star formation site and for LINER
excitation, respectively, but the high [N II]--to--H# ratio at the
inner edge of the ring and at the outer edge of the central concen­
tration needs further explanation. Buta et al. (1995) considered
the rings in NGC 7217 in detail and noted that the H#­ring co­
incides with the continuum (B­band) ring, but inside of both
there is a dust ring. Similar shifts between the stellar (gaseous)
and dust features are observed in spiral arms and large­scale
bars, and there they are attributed to shocks. Our high [N II]--
to--H# ratios in the center of NGC 7217 could be explained if
the circumnuclear gaseous ring has a radial velocity component
toward the center and if the central ionized gas disk is inclined
with respect to the plane populated by surrounding gas: then
shocks would be arisen in the required locations.
In our previous work (Zasov & Sil'chenko 1997) we re­
ported a turn of the isophote major axis and an increase of ellip­
ticity inside R = 1 ## using HST WFPC2 data obtained through
the F547M filter (the pivot wavelength is 5487 š
A). But dust,
which is obviously present in the center of NGC 7217, might
distort the true shape of the brightness (mass) distribution. Now
we add NICMOS data at 1.1 µm and 1.6 µm which are much less
affected by dust. A comparison of the NICMOS and WFPC2
data (Fig. 9) shows their consistensy: indeed, the isophote ma­
jor axis turns by # 90 # and the ellipticity increases in the very
center of the galaxy. The central gravitational potential shape in

O.K. Sil'chenko & V.L. Afanasiev: Decoupled nuclei and nuclear polar rings in regular spiral galaxies 487
Fig. 8. Isocontours of two­dimensional maps of the surface brightness
distributions for the emission lines [N ii]#6583 (top) and H# (middle)
in the center of NGC 7217: the gray­scaled background is the red
continuum distribution, in arbitrary units; the bottom panel presents
their intensity ratio (gray­scaled) overlaid by H# isophotes.
Fig. 9. Radial variations of the isophote morphological characteristics
in NGC 7217 according to the data obtained with the WFPC2 and
NICMOS of the Hubble Space Telescope.
NGC 7217 may be an ellipsoid with a polar orientation of the
projected largest axis, or an axisymmetric ellipsoid with the line
of nodes orthogonal to the global plane of the galaxy. The latter
suggestion is consistent with the alignment of the dynamical
major axis of the circumnuclear ionized gas: this coincidence
proves the circular rotation of the gas in the polar plane. The
radius of the polar gaseous disk, some 3 ## (250 pc), is larger
than the visible extent of the polar orientation of the photomet­
ric major axis, but as we see in the end of this section, the polar
stellar structure is more extended than it can be deduced from
the analysis of the integrated surface brightness alone -- a de­
composition of global structural components is necessary. We
would only note that the orientation of the dynamical major axis
of the stars within 3 ## from the nucleus implies a prolate form
of the central stellar structure.
5.2. The whole galaxy
Photometric analysis of the global structure of NGC 7217 has a
long history. Boroson (1981) tried to decompose the major axis
brightness profile in the B­band, but unsuccessfully, because
multiple blue rings made the B­band profile very irregular. Kent
(1986) analysed a 2D image in the r­band, more regular than
that in B, and decomposed it into a de Vaucouleurs' bulge and
an exponential disk, with a 1:3 luminosity ratio. According to
this decomposition, the exponential disk becomes the dominant
component at R # 20 ## . However, there are other, quite different
variants of decomposition. Buta et al. (1995) obtained a de Vau­
couleurs' bulge which dominates over the full radius range, with
a total bulge­to­disk luminosity ratio of 2.3--2.4. A similar result
was obtained by Baggett et al. (1998). Our analysis of the bright­

488 O.K. Sil'chenko & V.L. Afanasiev: Decoupled nuclei and nuclear polar rings in regular spiral galaxies
Fig. 10. Azimuthally averaged profiles
of the surface brightness of NGC 7217
in the I­band. The results of recurrent fit­
ting by exponential laws are presented at
every step for the residual brightnesses,
and the final result of fitting the initial
total profile by 3 exponential disks with
parameters given in Table 3 is shown in
the right bottom panel.
Table 3. Exponential parameters of the NGC 7217 image fits
Disk Radius range of fitting PA0 i µ0 , I­mag/## ## r0 , ## r0 , kpc
Outer 60 ## --110 ## 90 # 35 # 18.3 35.8 2.9
Intermediate 20 ## --50 ## 60 # 23 # 17.4 12.5 1.0
Inner 5 ## --20 ## 82 # 28 # 15.9 3.9 0.3
ness profile of NGC 7217 (Zasov & Sil'chenko 1997) revealed
the existence of two equally good decomposition variants: a
de Vaucouleurs' bulge plus an exponential disk, with the bulge
dominating over the full radius range, and a King's bulge plus
an exponential disk, with the bulge being fainter than the disk
in the radius range of 20 ## --180 ## . The scalelengths of the fitted
disks are very similar, 42 ## and 35 ## , respectively, but their cen­
tral brightness is of course very different. Moreover, even if one
considers the central part of the brightness profile only, where
the bulge predominance is beyond doubt, the effective radius
of the de Vaucouleurs' law diminishes when approaching the
center. Obviously, the surface brightness profile of NGC 7217
is too complex to be decomposed in unique way.
Since up to now there is no convergence in the pub­
lished results of the decomposition of the brightness profile of
NGC 7217, we would like to propose one more variant. For this
purpose we have used photometric data from the ING Archive
to calculate a brightness profile in the I­band. We have con­
verted it into the standard Cousins system by comparison with
the central I measurements from Sanchez­Portal et al. (2000).
We then applied a procedure of recurrent fitting, the results of
which are given in Table 3 and Fig. 10. First, the parameters
are determined for the outermost parts of the exponential disk,
the model map of this exponential disk is subtracted from the
original map, in the residual map the position angle of the ma­
jor axis and the mean ellipticity are measured, an azimuthally­
averaged profile of the residual brightness is calculated, and all
the steps above are repeated again. Fig. 10 presents the stages
of this recurrent fitting for NGC 7217. Curiously, at every step
the outer parts of the profiles look exponential. The innermost
of the three exponential profiles can be traced as close to the
nucleus as R = 5 ## . The scalelength of the third disk is 3. ## 86 or
0.3 kpc -- about 10% from the scalelength of the first disk. Such
a ratio is typical for a galaxy consisting of an exponential disk
and an exponential bulge (Courteau et al. 1996), though an ex­
ponential bulge in a Sab galaxy is a quite unusual phenomenon
(Andredakis et al. 1995, de Jong 1996, Moriondo et al. 1998).
However this fact allows us to treat NGC 7217 as a galaxy with
an exponential bulge and two exponential disks. The final resid­
ual map obtained by subtracting the three model exponential­
disk images from the original I­band image of the galaxy
(Fig. 11) is quite clear. One can see the stellar rings at R = 11 ##
and R # 30 ## , and these rings look asymmetrical: their southern
halves are brighter. For a face­on disk galaxy, such as NGC 7217,
it may mean either that the dust tori are thick, or that the dust
(gas) disk is inclined with respect to the stellar disk harboring

O.K. Sil'chenko & V.L. Afanasiev: Decoupled nuclei and nuclear polar rings in regular spiral galaxies 489
Fig. 11. The residual brightness map in the I­band after subtraction of
the best­fit three­disk model. The full dimensions of the field shown
are 100 ## â 100 ## . North is up, east is left.
the rings, or, even more generally, that the dust and stellar rings
have different inclinations because both can be inclined to the
stellar disk. And finally, in the very center of the residual bright­
ness map we see a feature which we expected to find: an oval
with a major semiaxis of 4 ## --5 ## elongated in the north­south
direction. This is just what we discussed at the end of the pre­
vious section: the stellar substructure corresponding to the area
of maximum stellar velocity dispersion (Fig. 3, right) and of the
enhanced #Fe# (Fig. 1, right).
6. Conclusions and discussion
We have found unique structure in the center of NGC 7217. The
continuum and emission­line isophotes within R # 3 ## (240 pc)
are elongated approximately in the north­south direction, and
the dynamical major axis of the ionized gas within R # 3 ##
also has PA 0 # -30 # ; and this coincidence shows a planar
circular character of the gas rotation. However, the plane of this
rotation is nearly orthogonal to the global plane of the galaxy
which has a line of nodes in PA 0 # 90 # . Interestingly, a similar
elongation in the north­south direction is seen in the contours
of the Lick #Fe# index distribution, whereas the magnesium­
line index Mg 2 decreases quite axisymmetrically with radius.
Therefore, we suggest the existence of a polar circumnuclear
stellar structure, which is iron overabundant with respect to the
solar abundance ratio.
Magnesium overabundant stellar systems are born when
the epoch of their star formation is shorter than one billion
years. Stellar systems with a solar magnesium--to--iron ratio are
thought to have long and continuous star formation. To obtain
iron overabundant stellar systems, chemical evolution modelists
have come up with a plausible scenario: the star formation must
be bursty, with pauses for a few Gyrs between the bursts. But
this suggests that in the central, polarly elongated stellar struc­
ture several, at least two, discrete star formation bursts have
occured. This can be related to the structure of the global disk
of the galaxy, which also consists of several (three) exponen­
tial segments with different slopes. Lin & Pringle (1987) have
shown that the exponential density profile of a stellar disk is a
natural product of dynamical evolution when the characteristic
time of the star formation is comparable to the viscous gas re­
distribution time, and its exponential scalelength is completely
determined by the initial radius of the gaseous disk. With this
idea in mind, we can propose the following scenario for the evo­
lution of NGC 7217: in the first stage, the primordial gaseous
disk was rather extended and a first stellar disk was formed
with a usual scalelength, some 3 kpc. But soon a catastrophique
event occured, throwing the gas suddenly toward the center by
external interaction or internal instability, and the formation of
the next part of the disk resulted in a smaller scalelength. The
cause of this catastrophique event cannot be diagnosed in the
present epoch: though the galaxy looks isolated now, it may
have had small satellites some Gyrs ago which may be fully
accreted after playing a role of external disturber. After a few
Gyrs of quiescent star formation in the global disk of NGC 7217
a second catastrophique event occured, again gas flows into the
center, and the exponential scalelength of the next, third gen­
eration component of the stellar disk (bulge?) is even smaller
than that of the second. Every abrupt re­distribution of the gas
along the radius had to be accompanied by a star formation burst
in the center; if the temporal separation of two bursts was two
to three Gyrs, this would result in iron overabundance in the
central stellar concentration.
But if the central gaseous and stellar structures were formed
from the gas of the global disk, why does the circumnuclear
gas now rotate in the polar plane? This may be a consequence
of the intrinsic dynamical evolution of the viscous subsystem.
Sofue & Wakamatsu (1994) proposed that gas rotating in the
plane with non­axisymmetric perturbation (e.g., a bar) passes
through shocks at the edges of the bar and loses its tangential
momentum first of all; it then flows into the center and there
loses its radial velocity component, and finally only its verti­
cal velocity remains unaffected. This would result in the oc­
curence of circumnuclear gaseous polar rings (disks) in barred
galaxies. Anantharamaiah & Goss (1996) reported such a ring
in NGC 253, and we have found several other examples: in the
Sb galaxy NGC 2841 (Sil'chenko et al. 1997b) and in the S0/a
galaxy NGC 6340 (Sil'chenko 2000) the presence of a compact
circumnuclear gaseous polar disks may be caused by the triax­
iality of their bulges; in NGC 2841 we directly observe large­
scale shocks at R = 1 - 2 kpc (Afanasiev & Sil'chenko 1999).
In the S0 galaxy NGC 7280, which also possesses a circumnu­
clear gaseous polar disk, we also detected an intermediate­scale
high­contrast bar with an asymmetric dust lane -- a true signa­
ture of a shock (Afanasiev & Sil'chenko 2000). In NGC 7217
the clear evidence for the present or past existence of a bar is

490 O.K. Sil'chenko & V.L. Afanasiev: Decoupled nuclei and nuclear polar rings in regular spiral galaxies
its three rings. Buta et al. (1995) have found a low­contrast bar
(more exactly, a triaxial structure) in PA # 60 # by applying a
Fourier analysis to a deep I­band image and showed that even
such a weak perturbation may produce rings. Perhaps, this bar
was stronger earlier: mass accumulation in the center had to
provoke its dissolution, making it only a low­contrast feature
in the present epoch. At the time when it was stronger, it could
have stimulated gas inflow, accumulating it in the polar plane
around the nucleus.
Acknowledgements. We thank the post­graduate student of the Spe­
cial Astrophysical Observatory A. V. Moiseev for supporting the ob­
servations at the 6m telescope. The 6m telescope is operated under
the financial support of Science Ministry of Russia (registration num­
ber 01­43). During the data analysis we have used the Lyon­Meudon
Extragalactic Database (LEDA) supplied by the LEDA team at the
CRAL­Observatoire de Lyon (France) and the NASA/IPAC Extra­
galactic Database (NED) which is operated by the Jet Propulsion Lab­
oratory, California Institute of Technology, under contract with the Na­
tional Aeronautics and Space Administration. This research has made
use of the ING Archive. The WHT and JKT telescopes are operated
on the island of La Palma by the Royal Greenwich Observatory in the
Spanish Observatorio del Roque de los Muchachos of the Instituto de
Astrofisica de Canarias. The research is also partly based on obser­
vations made with the NASA/ESA Hubble Space Telescope, obtained
from the data archive at the Space Telescope Science Institute, which is
operated by the Association of Universities for Research in Astronomy,
Inc., under NASA contract NAS 5­26555. The work was supported by
grant 98­02­16196 of the Russian Foundation for Basic Researches and
by the Russian State Scientific­Technical Program ``Astronomy. Basic
Space Researches'' (the ``Astronomy'' section).
References
Afanasiev V.L., Sil'chenko O.K., 1999, AJ 117, 1725
Afanasiev V.L., Sil'chenko O.K., 2000, AJ 119, 126
Afanasiev V.L., Vlasyuk V.V., Dodonov S.N., Sil'chenko O.K., 1990,
Preprint SAO N54, Special Astrophys. Obs., Nizhnij Arkhyz
Afanasiev V.L., Dodonov S.N., Drabek S.V., Vlasyuk V.V., 1996,
MPFS Manual. SAO Publ., Nizhnij Arkhyz
Anantharamaiah K.R., Goss W.M., 1996, ApJ 466, L13
Andredakis Y.C., Peletier R.F., Balcells M., 1995, MNRAS 275, 874
Athanassoula E., 1996, In: Buta R., Crocker D.A., Elmegreen B.G.
(eds.) Barred Galaxies. ASP Series 91, p. 309
Baggett W.E., Baggett S.M., Anderson K.S.J., 1998, AJ 116, 1626
Block D.L., Elmegreen B.G., Wainscoat R.J., 1996, Nat 381, 674
Boroson T., 1981, ApJS 46, 177
Buta R., van Driel W., Braine J., et al., 1995, ApJ 450, 593
Cardiel N., Gorgas J., Cenarro J., Gonzalez J.J., 1998, A&AS 127, 597
Courteau S., de Jong R.S., Broeils A.H., 1996, ApJ 457, L73
de Jong R.S., 1996, A&A 313, 45
Dressler A., 1984, ApJ 286, 97
Elmegreen D.M., Chromey F.R., Bissell B.A., Corrado K., 1999, AJ
118, 2618
Emsellem E., Bacon R., Monnet G., Poulain P., 1996, A&A 312, 777
Gilmore G., Wyse R.F.G., 1991, ApJ 367, L55
Goudfrooij P., Emsellem E., 1996, A&A 306, L45
Jablonka P., Martin P., Arimoto N., 1996, AJ 112, 1415
Jore K.P., Broeils A.H., Haynes M.P., 1996, AJ 112, 438
Kent S.M., 1986, AJ 91, 1301
Laine S., Knapen J.H., Perez­Ramirez D., Doyon R., Nadeau D., 1999,
MNRAS 302, L33
Lin D.N.C., Pringle J.E., 1987, ApJ 320, L87
Marconi G., Matteucci F., Tosi M., 1994, MNRAS 270, 35
Matteucci F., 1994, A&A 288, 57
Merrifield M.R., Kuijken K., 1994, ApJ 432, 575
Moriondo G., Giovanardi C., Hunt L.K., 1998, A&AS 130, 81
Peterson C.J., Rubin V.C., Ford W.K., Roberts M.S., 1978, ApJ 226,
770
Pogge R.W., 1989, ApJS 71, 433
Rubin V.C., Burstein D., Ford W.K., Thonnard N., 1985, ApJ 289, 81
Ryder S.D., Zasov A.V., McIntyre V.J., Walsh W., Sil'chenko O.K.,
1998, MNRAS 293, 411
Sanchez­Portal M., Diaz A.I., Terlevich R., et al., 2000, MNRAS 312,
2
Sil'chenko O.K., 1993, Pis'ma v Astron. Zh. 19, 701
Sil'chenko O.K., 1999, Pis'ma v Astron. Zh. 25, 176
Sil'chenko O.K., 2000, AJ 120, 741
Sil'chenko O.K., Afanasiev V.L., Vlasyuk V.V., 1992, AZh 69, 1121
Sil'chenko O.K., Burenkov A.N., Vlasyuk V.V., 1997a, New Astron­
omy 3, 15
Sil'chenko O.K., Vlasyuk V.V., Burenkov A.N., 1997b, A&A 326, 941
Sofue Y., Wakamatsu K.­I., 1994, AJ 107, 1018
Tantalo R., Chiosi C., Bressan A., 1998, A&A 333, 419
Thornley M.D., Mundy L.G., 1997, ApJ 484, 202
Trager S.C., Worthey G., Faber S.M., Burstein D., Gonzalez J.J., 1998,
ApJS 116, 1
Vlasyuk V.V., 1993, Astrofiz. issled. (Izv. SAO RAS) 36, 107
Whitemore B.C., Schechter P.L., Kirshner R.P., 1979, ApJ 234, 68
Worthey G., 1994, ApJS 95, 107
Worthey G., Faber S.M., Gonzalez J.J., 1992, ApJ 398, 69
Worthey G., Faber S.M., Gonzalez J.J., Burstein D., 1994, ApJS 94,
687
Zasov A.V., Sil'chenko O.K., 1997, AZh 74, 824