VLA HI imaging of the brightest spiral galaxies in Coma

34 Pages·0.77 MB·English

by  H. Bravo-Alfaro

VLA HI imaging of the brightest
spiral galaxies in Coma
H. Bravo-Alfaro
Observatoire de Paris DAEC, and UMR 8631, associ´e au CNRS et `a l’Universit´e Paris 7, 92195
0
0 Meudon Cedex, France.
0
2 Departamento de Astronom´ıa, Universidad de Guanajuato. Apdo. Postal 144 Guanajuato 36000.
n M´exico.
a
J

1
1 V.Cayatte

2 Observatoire de Paris DAEC, and UMR 8631, associ´e au CNRS et `a l’Universit´e Paris 7, 92195
v
Meudon Cedex, France.
5
0
4
2
J. H. van Gorkom
1
9
Department of Astronomy, Columbia University, New York, New York 10027.
9
/
h and
p
-
o
r
t C. Balkowski
s
a
: Observatoire de Paris DAEC, and UMR 8631, associ´e au CNRS et `a l’Universit´e Paris 7, 92195
v
i Meudon Cedex, France.
X
r
a
ABSTRACT
We have obtained21 cm imagesof 19 spiral galaxiesin the Comaclus-
ter, using the VLA in its C and D configurations. The sample selection
was based on morphology, brightness, and optical diameters of galaxies
◦ i
within one Abell radius (1.2 ). The H detected, yet deficient galaxies
i
show a strong correlation in their H properties with projected distance
i >
from the cluster center. The most strongly H deficient (Def 0.4)
HI
galaxies are located inside a radius of 30′ (∼0.6 Mpc) from the center of
1
Coma, roughly the extent of the central X-ray emission. These central
i
galaxies show clear asymmetries in their H distribution and/or shifts
between the optical and 21 cm positions. Another 12 spirals were not
detected in Hi with typical Hi mass upper limits of 108 M . Seven out
⊙
of the twelve non detections are located in the central region of Coma,
′
roughly within 30 from the center. The other non detections are to the
east and southwest of the center.
i
Seven so calledblue disk galaxiesin Coma were observedin H and six
were detected. These galaxies are relatively close to the central region
of Coma. The non detected one is the closest to the center. The six
i
detected blue galaxies are mildly H deficient. We did a more sensitive
i
search for H from 11 of the 15 known post starburst galaxies in Coma.
None were detected with typical Hi mass limits between 3 and 7×107
M .
⊙
Our results present and enhance a picture already familiar for well
i
studied clusters. H poor galaxies (deficient ones and non-detections)
i
are concentrated toward the center of the cluster. The H morphology
i
of the central galaxies, with optical disks extending beyond the H disks
is unique to cluster environments and strongly suggests an interaction
with the IGM. A new result in Coma is the clumpy distribution of gas
deficiency. In the cluster center the deficient galaxies are to the east
while the non-detections are to the west. In the outer parts the gas
rich galaxies are north of Coma, non-detected spirals are found in the
NGC 4944 group to the east and NGC 4839 group to the south west.
This supports recent findings that merging of groups is ongoing in the
center of Coma, further out the NGC 4944 and NGC 4839 must have
passed at least once through the core, while the galaxies to the north
have yet to fall in.
Subject headings: Galaxies: Clusters: Individual: Coma
2
1. Introduction. al. 1990, 1994). In the center of Virgo the
Hi disks are much smaller than the optical
Two of the outstanding questions on clus- disks. The actual Hi removal mechanisms at
ter of galaxies are: what causes the density
work in the center of Virgo could be identi-
morphology relation and how dynamically re- fied by a detailed comparison of the Hi sur-
laxed are clusters. For the density morphol-
face density distribution of the cluster galax-
ogy relation the question is: are all galax-
ies with their counterparts in the field (Cay-
ies formed by the same process and do they atte et al. 1994). Thus, the Hi distribution
evolve differently in different environments?
of individual galaxies reflects the effect of the
Ordoestheenvironment playadominant role
cluster environment, while the location and
at the time of formation? On a larger scale
velocities of gas-rich versus gas-poor galaxies
the question is whether clusters at low red-
can help identify possible substructure in the
shift are dynamically relaxed or whether they
clusters (Valluri et al. 1999). Gas can also
arestillaccreting significant amountsofmass.
be used to probe the orbital history of the
The answer to this question may contain a
galaxies. More recently other clusters have
clue to the density of the universe (Thomas been imaged in Hi with the VLA and WSRT:
et al. 1998). Observations of clusters at inter-
Hydra (McMahon 1993), Ursa Majoris (Ver-
mediate redshift suggest that clusters evolve
heijen 1996), A 2670 (van Gorkom 1996),
rapidly. Already at intermediate redshifts
A 262 (Bravo–Alfaro et al. 1997), and Her-
they differ markedly from clusters at low red-
cules (Dickey 1997). Preliminary results on
shifts. Many clusters at redshifts larger than
Coma were obtained by Sullivan (1989) in
z∼0.3areknowntocontainalargepopulation
the early days of the VLA. In spite of poor
of blue galaxies, indicative of enhanced star
sensitivity and interference problems in the
formation, which is not found in similar clus- data, he observed disturbed Hi disks for a few
ters at z< 0.1 (Butcher & Oemler 1978, 1984;
galaxies in the center.
Dressler & Gunn 1992, Poggianti et al. 1999,
In this paper we present a more complete
Dressler et al. 1999). Secondly these clusters
and much more sensitive study of the Coma
oftencontaingalaxieswhich haverecently un-
cluster. We expect the environmental effects
dergone a starburst, but which are not cur-
in Coma to be stronger than in Virgo, be-
rently forming stars, displaying spectra with
cause Coma is a much richer cluster and be-
strong Balmer absorption lines but no emis-
cause its central X–ray emission is almost
sion lines at all. These galaxies are commonly
six times more luminous than in Virgo. In-
referred to as “post-starburst” (PSB) galax-
deed, single dish observations have shown
ies.
thatComaisoneofthemostHideficientclus-
WeaddressbothquestionswithHiimaging ters; extremely Hi deficient galaxies are seen
studies of nearby clusters, which are directly out to 1.5 rA (the Abell radius, rA=1.7′/z,
comparable to more distant ones. Detailed
where z is the redshift), while in other clus-
Hi imaging of the nearest cluster of galax- ters strongly Hi deficient galaxies are gen-
ies, Virgo, has shown that the neutral gas
erally not seen beyond 0.75 rA (Bothun et
component gets affected dramatically by the
al.1985; Giovanelli andHaynes 1985; Gavazzi
hotintergalacticmedium(IGM)(vanGorkom
1987; 1989). The Coma cluster is in fact the
et al. 1984, Warmels 1988, and Cayatte et
3
richest of the nearby clusters (richness class ter). Three fields (1, 2, and 9) are centered
1
2, and z = 0.023 ), and represents the clos- around the core, covering most of the Coma
est equivalent to what is seen at intermediate velocity range. Field 8 was pointed south of
redshifts, as it has blue disk galaxies (Bothun the elliptical NGC 4839, which is the dom-
& Dressler 1986, and references therein) and inant galaxy of the SW group. Five other
PSB galaxies (Caldwell, Rose, & Dendy 1999, fields (3, 4, 6, 7, and 10) contain the blue
and references therein), which are commonly disk galaxies reported by Bothun & Dressler
seenindistantclusters. ThismakesComathe (1986) and most of the PSBs reported by
perfect link between nearby and more distant Caldwell et al. (1993) and Caldwell & Rose
clusters. (1997). The remaining three fields (5, 11, and
In the present paper we discuss the global 12) were centered in the outer regions of the
properties of Coma, on the basis of our Hi re- cluster, where nostrong environmental effects
sults, combined with recent observations on are expected, providing a comparison sample.
different wavelengths and numerical work. A The VLA survey of Coma was carried out
detailed analysis of individual galaxies and in C and D array, during three runs, in March
morespecificregionsofComa,willbethesub- 1995, April 1996, and March 1999. In March
ject of a forthcoming paper (Bravo–Alfaro et 1995, an exploratory observation was done of
al. 1999). The plan of this paper is as fol- fields 2, 4 and 5, in D array. Twelve fields
lows: Observations and data reduction are wereobserved inCarrayinApril1996. Fields
described in Section 2. Observational results 1 and 10 were reobserved, March 1999, to get
are given in Section 3. In Section 4 we discuss very sensitive observations of the PSB galax-
what the observations have taught us about ies in these fields. For the fields observed in
Coma at large. Conclusions are summarized both, C and D configurations (fields 1, 2, 4, 5,
in Section 5. A description of Hi properties and 10), the data were combined to improve
of individual galaxies is given in Appendix A. sensitivity. The combined data provide the
higher spatial resolution and a lower veloc-
2. Observations ity resolution. In the D array observations of
1995 online Hanning smoothing was used, af-
We selected the 25 brightest spiral galax-
ter which every other channel was discarded.
ies in Coma with morphological type later
This leaves a set of 31 independent channels
than S0. All those galaxies have B magnitude
with a velocity spacing of 43 km/s. In C ar-
smaller than 15.7 mag and optical diameters
ray no online Hanning smoothing was used,
larger than 0.5′. Fig. 1 shows the location of
providing a set of 63 channels with a velocity
2
the 12 fields observed with the VLA, up to
resolution of 21.7 km/s. Most of these obser-
1.2◦ (1 rA) from the center of Coma (we take vationswere afterwardsHanning smoothedto
the position of NGC 4874 as the Coma cen-
the same velocity resolution as the D array
data. The C array data were tapered to an
1We assume a velocity of 7000 km s−1, and H◦ = 100
angular resolution of 30 arcsec, and, in most
km s−1 Mpc−1 throughout this paper.
2The National Radio Astronomy Observatory is oper- cases, we obtained a resolution of 35 arcsec
for the combined C+D data.
ated by Associated Universities, Inc., under coopera-
tiveagreementwiththeNationalScienceFoundation. StandardVLAcalibrationandimagingpro-
4
cedures were applied, using the NRAO’s as- maps.
tronomical image processing system (AIPS).
Data cubes were made using nearly pure nat- 3. Results
ural weighting to obtain a higher sensitivity.
We detected 19 galaxies in this survey; 17
A set of eight line–free channels on either
spirals, one irregular, and one interacting sys-
side of the band was used (four channels in
tem. Table 2 lists those galaxies with their
the case of C+D combined data) to define
observational parameters. Columns 1 and 2
a mean continuum image. This image was
givethegalaxyidentification, Column3and4
then subtracted from the channel maps form-
ing an Hi line emission cube. We used this the R.A. and Dec. (1950) from the NED data
base, except (1) taken from the LEDA data
cube to search for 21 cm line emission and
base, Column 5 gives the VLA configuration,
to determine the spatial position and veloc-
Column 6 the synthesized beam size (in arc-
ity range for each signal. Finally, a new con-
sec), Column 7 gives the morphological type
tinuum image was made using only line–free
from Dressler 1980, except when indicated:
channels. The channels containing line emis-
(1) from Huchra et al. 1990, (2) from the
sion were CLEANed (for more details on data
LEDA database, and (*) for uncertain classi-
reductionsee Bravo–Alfaro1997). Thermsin
fication. Column 8 gives the blue total mag-
thefinalcubesistypically0.37mJy/beamper
nitudeobtainedfromtheLEDAdatabase, ex-
channel. The observationshave onaveragean
Hi mass detection threshold of 108 M , corre- cept (1) taken from the NED database; Col-
⊙
sponding to 6×σ×21.7km sec−1 (the channel umn 9 gives the mUV-b color, from Donas et
al. 1995.
width) and a typical surface brightness sensi-
tivity of 2 to 4×1019 cm−2 (corresponding to In Table 3 we list the Hi parameters de-
2.5 rms). Fields 1 and 10, were reobserved in rived from the observations: Columns 1 and
1999 for 20 and 15 hours respectively; there 2 are as in Table 2, Column 3 gives the ob-
our detection thresholds are lower: the rms served field, Column 4 and Column 5 give the
per channel is around 0.20 mJy/beam, and central Hi velocity and the velocity width, re-
the Hi mass detection threshold is 2.4×107 spectively. As thesignal-to-noiseratio is poor
M in the center of the field. for some galaxies, we use as a homogeneous
⊙
criterium to obtain the central Hi velocity the
The observational parameters are listed in
central channel displaying emission. The un-
Table 1, where Column 1 indicates the ob-
certainty in this value is around half the ve-
served field, Columns 2 and 3 give R.A. and
locity resolution, i.e. ∼11 km s−1. The ve-
Dec. (1950) of each pointing, Column 4 gives
locity width is defined as given by the range
the VLA configuration, Column 5 the inte-
of channels containing Hi emission. The error
gration time, Column 6 the total bandwidth,
in this case is the velocity resolution, around
Column 7 the heliocentric velocity of the cen-
22 km s−1. Column 6 gives the Hi total flux
tral channel, Column 8 the channel separa-
tion (in km s−1), Column 9 the rms noise per corrected for the primary beam attenuation,
with the corresponding error. Column 7 gives
channel after the continuum subtraction, and
the continuum intensity, Column 8 the total
Column 10 the flux density per beam area (in
Hi mass, Column 9 and 10 give the Hi de-
mJy/beam)equivalent to1.0Kinthechannel
ficiency and the projected distance from the
5
cluster center respectively. cal type taken from Dressler 1980, except (1)
The main result of this paper is summa- taken from Huchra et al. 1990, (2) from the
rized in Fig. 2. It shows a synthetic view NED database, and (3) from the RC3 cata-
of all the galaxies detected in Hi. They are log; (*) means uncertain classification. Col-
placed at their correct location in the cluster umn 7 gives the optical velocity taken from
and magnified by a factor 7. Contours of the the LEDA database, Column 8 the rms noise
individual Hi images (the first contour corre- per channel corrected for the primary beam
sponds to a column density of 3×1019 cm−2) response at the position of the galaxy. The
are overlaid on optical images shown in grey- Hi mass upper limit given in Column 9 cor-
scale (DSS). The cross indicates the center responds to 6 times the rms noise, multiplied
of the cluster (taken to be the position of by the channel width (typically 21.7 km s−1).
NGC 4874), and the central contours draw Column 10 gives the lower limit to the Hi de-
the X-ray emission in the 0.5–2 keV energy ficiency.
band, as observed with ROSAT (Vikhlinin, In addition, we made a special effort to de-
et al. 1997). This figure displays a wealth of tectthesocalledabnormalspectrumgalaxies,
information. As in Virgo, the first thing to reported in Coma by Caldwell et al. (1993)
note is that the Hi disks in the outer parts and Caldwell and Rose (1997). Some of
of the cluster are much larger than the opti- them display ongoing star formation activ-
cal disks. But something different in Coma, ity (SB), and others a recent peak of star
are the asymmetries and even displacements formation (PSB). Most of these galaxies are
of the shrunken Hi disks seen in the center. found southwest of the cluster center (see
All the shrunken HI disks are in projection Sect. 3.4.2). In Table 5 we give a list of all the
within the boundaries of the X–ray emission peculiar spectrum galaxies observed in this
(see Fig. 2), suggesting that an interaction survey; none was detected in Hi. Column
with the IGM may be at work. Near the 1 and 2 give the galaxy identification (the
cluster center, the distribution of Hi detected first corresponds to the Dressler, 1980 cat-
galaxies is very non uniform: most of the de- alog number), Column 3 gives the observed
tections lie east of the center, and there are field, Column 4 and 5 the (1950) R.A. and
almost no detections west of the center. In Dec. Column 6 the morphological type, and
the zone between the center of Coma and the Column 7 the heliocentric velocity, both from
SW group, only one galaxy (Mrk 058), with the NED database. Column 8 gives the rms
very low gas content, was detected. noise per channel corrected for the primary
Twelve galaxies from the original observed beam response at the position of the galaxy.
sample (morphological type later than S0 and Column 9 gives the Hi mass upper limit for
mB <15.7)were not detected (Table 4). They both, SBandPSBnondetected galaxies, typ-
are within 20′ from the center of the ob- ically between 3 and 7×107 M⊙. We did not
served fields and within the observed veloc- estimate the Hideficiency parameter, as most
ity range. Table 4 gives in Columns 1 and of these galaxies are lenticular (morphological
2 the identification, Columns 3 and 4 the type S0).
R.A. and Dec. (1950), in Column 5 the ob-
3.1. The HI Content
served field, Column 6 gives the morphologi-
6
3.1.1. Comparison with previous works sured by Gavazzi. This galaxy is very close to
the detected Hi rich galaxy KUG 1255+275,
Five of the 19 detected galaxies are new
both spatially (4.6′) and in velocity, thus this
Hi detections. Three of them (FOCA 195,
result may be due to confusion inside the
KUG 1255+275, and KUG 1258+287) are
Arecibo beam. On the other hand, the de-
Hi rich systems with relatively low optical
tection of NGC 4944 is reported by Gavazzi
surface brightness, which have not been ob-
(1987) as a marginal one, and it was not con-
served previously in single dish Hi surveys.
firmed in later surveys (e.g. Gavazzi 1989).
NGC 4907 and CGCG 160–106, were not
detected in previous single–dish Hi surveys
3.1.2. The HI deficiency
probably because of their low total Hi flux
(seeTable3). CGCG160–106wasalreadyde- To quantify the Hi content of the spirals
tected with the VLAby Sullivan (1981). Four as compared to so called field spirals we use
of the galaxies detected in the present survey the Hi deficiency parameter (DefHI), follow-
(NGC 4907, CGCG 160–106, Mrk 058, and ing the definition by Giovanelli & Haynes
IC 4040), were not detected in the most re- (1985, and references therein), where the de-
cent survey carried out with the Arecibo tele- ficiency parameter is the log of the ratio of
scope (Haynes et al. 1997), either because of the average Hi mass of isolated spirals of the
low Hi flux, or because several objects were same morphological type and the observed
found inside the beam. Hi mass. For our sample, we use the mor-
phologicaltypesfromDressler (1980),orfrom
In Fig. 3 we show a comparison between
Huchraetal.(1990)whenthefirstisnotavail-
single dish and VLA measurements. There is
able. We should remark that the Hi deficien-
ingeneralagoodagreement, showing thatthe
cies in Coma are less well defined than for
VLA has not missed any extended flux. Two
more nearby clusters due to considerable un-
discrepant cases are seen in Fig. 3, NGC 4848
certainty in the morphological classification.
and IC 842. We obtained for the former, sys-
tematicly a lower totalHiflux thanthevalues A diagram of the Hi deficiency versus pro-
previously reported (Chincarini et al. 1983, jected distance to the cluster center is shown
Giovanelli & Haynes 1985, Gavazzi 1989 and in Fig. 4, where all detected galaxies are
Haynes et al. 1997). We confirm the asym- shown with black circles and lower limits
metric distribution found by Gavazzi (1989) for non detections (only galaxies from Ta-
but we probably miss some extended Hi emis- ble 4 are included in this figure) with tri-
sion. We obtained for IC 842 a slightly higher angles. The merging system NGC 4922 is
Hi flux (∼30%) than previous single dish ob- not plotted in this figure because of its pe-
servations (Chincarini et al. 1983, Bothun culiarity (see Appendix A). Clearly the most
et al. 1985, Gavazzi 1989, and Haynes et Hi deficient galaxies are closer to the cluster
al. 1997). core. All strongly Hi deficient detected galax-
ies (Def>0.4) are inside a projected radius of
We do not confirm two detections pre-
viously reported, Mrk 056 and NGC 4944, 30′ (0.4rA) from the center of Coma. For the
non detected galaxies the trend of deficiency
whichwereonlymarginallydetectedbyGavazzi
(1987). Our Hi flux upper limit for Mrk 056 with the projected distance to the center is
not as clear as it is for the detected ones, but
(CGCG 160–064) is well below the flux mea-
7
7 of the non-detections areprojected inside or nitude of 15.7, with morphological type later
very close to the X–ray emission (UGC 8071, than S0 (Table 4). We also show in Fig. 5
NGC 4851, IC 3943, NGC 4858, IC 3955, the contours of the X–ray emission (ROSAT)
KUG 1258+277, and CGCG 160–261), as centered on Coma, as reported by Vikhlinin
shown in Fig. 5. et al.(1997). The centralregionofComa con-
For the non detected galaxies we used 6 tainsmostofthestronglyHideficientgalaxies
times the r.m.s. multiplied by the veloc- in the cluster, including the very bright early
ity channel width to calculate both, the de- spiralsNGC4911,NGC4921,andNGC4907,
ficiency and the Hi mass upper limit shown all of them very Hi deficient. The giant spi-
in Table 4. This value is estimated on the ral NGC 4911, detected in X-ray, is thought
basis of our detection threshold. Another to be the dominant galaxy of a group cross-
method to estimate the Hi mass upper limit ing the main Coma body. NGC 4921 (clas-
for non detected galaxies could consider the sified as “anemic” by van den Bergh, 1976)
expected velocity width for a big spiral, e.g. and NGC 4907 show very perturbed Hi maps;
300 km s−1, rather than the observed chan- they have large velocities relative to the mean
nel width. In this fashion the total flux upper cluster velocity (1179 and 1521 km s−1 re-
limitwouldbeobtainedasr.m.s.×2.5×300 kms−1. spectively). Fourothergalaxiesweredetected
But as these galaxies are not in a typical en- inside a radius of 30′ (∼0.60 Mpc) which is
vironment we do not know at all what the roughly the area defined by the X–ray emis-
Hi velocity width could be after a stripping sion shown in Fig. 5. With no exception these
process. Using this different limit does not galaxies show deeply perturbed Hi distribu-
substantially change the plot in Fig. 4. tions, some of them smaller than the corre-
sponding optical disks (see Fig. 2).
Although a possible correlation with dis-
tance to cluster center can be considerably
3.3. The SW of Coma
weakened by projection effects (Chamaraux
et al. 1980), the fact that the most deficient The presence of a group in the SW of
galaxies are almost all situated within the Coma, associated with the giant elliptical
central and southwestern X–ray emission sug- galaxy NGC 4839, has been well established
gests that an interaction with the IGM is at through optical and X–ray studies (e.g. Briel
work. More convincing evidence for this can et al. 1992, White et al. 1993, Colles & Dunn
be obtained by looking at the relative sizes 1996, Biviano et al. 1996). The group sits
of the Hi disks. No other mechanism would 40′ SW of the cluster core, and coincides with
cause the Hi disks to be smaller than the op- the second most intense X–ray peak. This
tical ones. We will address these issues in a group represents 6% of the total X–ray emis-
following paper. sion, and contains about 10% of the mass of
the cluster. The question whether this group
3.2. The central region is falling into the center of Coma (Colles &
Dunn 1996), or has already passed following
Fig. 5 shows the position of the detections
a straight path (Burns et al. 1994), is still
with crosses, andthe12 nondetected galaxies
a matter of debate. Recently, Caldwell &
with triangles. We consider here as non de-
Rose (1997) concluded that the SW group
tections those galaxies brighter than a mag-
8
has already passed through the main body 3.4.1. The blue disk galaxies
of Coma, based on the velocity structure of
As mentioned before Coma contains blue
the group and the presence of PSB galaxies
disk and post–starburst galaxies, similar to
in that zone. Our field number 8, which in-
those seen at intermediate redshifts. Bothun
cludes NGC 4839, and field number 10 in the
& Dressler (1986) obtained spectroscopy for
zone between this galaxy and the cluster cen-
seven blue galaxies in Coma, andshowed that
ter, try to clarify this controversy.
they had not only experienced a star burst,
Of the 18 galaxies surrounding NGC 4839
but were still forming stars at a high rate.
and belonging to the group (Biviano 1998),
Considering various mechanisms to produce
four were observed and none were detected.
the starburst phenomena these authors con-
This is perhaps not too surprising since all of
clude with a mild preference for ram pres-
them are classified as S0. We did however de-
sure induced star formation. The main rea-
tect three late type galaxies just south of the
son for rejecting galaxy–galaxy induced star
NGC 4839 group (IC 3913, KUG 1255+275
formation is that the blue disks do not ap-
and Mrk 057), just outside the X–ray emis-
pear to be interacting (except the non de-
sion(Figs.A2andA5). Theirradialvelocities
tected NGC 4858, possibly interacting with
are close to the systemic velocity of the group
NGC 4860 and member of an aggregate, Con-
(7339±329 km s−1, Colless & Dunn 1996),
selice and Gallagher 1998). All seven blue
thus they are likely to be group members.
disk galaxies were observed in this survey
These galaxies have undisturbed morpholo-
and six of them were detected: NGC 4848,
gies and a normal Hi content, which makes
Mrk058,CGCG160–086,IC4040,CGCG160–
it extremely unlikely that they have crossed
098 and NGC 4926-A. The first four are pro-
the cluster core.
jected inside the X–ray emission (Fig. 6) and
Our Hi results are inconclusive. The non are Hi deficient (DefHI> 0.5). They display
detections in the close vicinity of NGC 4839, very perturbed Hi distributions. The Hi de-
andthepresenceofseveralSBandPSBgalax- ficiencies and gas distribution strongly sug-
iesinthatzonewithaverylowHicontent(see gest that these galaxies are being ram pres-
Sect. 3.4.2 and Table 5), support the hypoth- sure swept by the IGM.
esis that the group has gone through. The The locations of the Hi detected blue disk
3 gas rich galaxies pose a problem. If they
galaxies are indicated with crosses in Fig. 6.
are members of the group, the group cannot
Most of them are in an annular region be-
have passed through the main cluster. Al- tween 20′ and 30′ from the cluster center,
ternatively they may be new members of the
which coincides with the outermost contour
group, only recently accreted after the group
of the X–ray emission (corresponding to a
has gone through the cluster center. gas density of ∼3×10−4 cm−3). This is simi-
lar to what has been seen in Butcher–Oemler
3.4. HI content of SB and PSB galax-
clusters, where blue galaxies seem preferen-
ies in Coma.
tially located in an annular region outside
the cluster core (e.g. Butcher and Oemler
1984; Mellier et al. 1988). Dressler & Gunn
(1990) reported that star forming galaxies
9
seem to avoid the cluster core, appearing first The spectra indicate that a burst of star for-
at a radius of ∼0.5 Mpc (0.5 Mpc=25′ in mation was recently truncated (∼1 Gyr ago).
Coma). Furthermore, theUVsurvey ofComa Among the 15 PSBs 8 are in the SW region,
by Donas et al. (1995) revealed that 38% of the remaining are near the center and in the
the UV flux is produced in a ring lying be- NE (Caldwell et al., 1993). We observed 11
tween 20′ and 30′ of the cluster center. This of those galaxies, and detected none (see Ta-
suggeststhatthiseffectmustberelatedtothe ble 5). The position of those galaxies are
global properties of the cluster, perhaps IGM shown in Fig. 6 with dotted circles. The X-
induced star formation can best explain the ray contours (Vikhlinin et al. 1997) are also
shell of blue starburst seen in Coma and sim- drawn.
ilarly in high redshift clusters (Oemler 1992). The NGC 4839 group has an average ve-
We also show in Fig. 6, with triangles, the locity close to the peak in the PSBs veloc-
non-detected blue disk galaxy (NGC 4858) ity distribution (Biviano et al. 1996). This
and three starburst like objects reported by suggests that the PSBs could be part of the
Caldwell et al.(1993),which were alsonot de- group and they have been stripped of their
tected in this survey (Table 5). The blue disk gaswhen they passed throughthe Coma core.
is projected well inside the X–ray emission The few PSBs in the center and NE of Coma
and star formation followed by ram pressure could be old members of the SW group, pre-
stripping may have exhausted their Hi reser- viously pulled out from the group when it
voir. The other three galaxies show abnor- passed through the cluster center (Caldwell
mal spectra with strong emission lines (Cald- et al. 1993).
well et al. 1993). These objects are also near Our results suggest an evolutionary se-
the edge of X–ray emission, in the SW direc- quence, where galaxies first become blue be-
tion. Theirnondetectionsuggeststhatinthis cause of IGM induced star formation. Star
zone, between the center and the SW group, formation and further ram pressure stripping
the gas is more easily removed. Note that makes them Hi deficient. The displaced and
starbursting galaxies in other directions (e.g. shrunken Hi disks of the blue disk galaxies in-
NGC 4848, IC 4040, and CGCG 160–086), at dicate that some interaction with the IGM is
similardistancesfromthecenter, aredetected indeed going on. The next step in this evo-
in Hi. lutionary sequence is the PSB phase, where
galaxies have lost most of the Hi gas, and as
3.4.2. The post starburst galaxies
consequence, star formation stops on a rela-
tively short time scale. In this picture all the
Caldwell et al. (1993)and Caldwell & Rose
PSBs would be in an advanced stage of gas
(1997) reported a total of 22 abnormal spec-
stripping.
trum galaxies in Coma. Five of them dis-
play Balmer absorption and emission lines
3.4.3. Correlation between Colors and HI–
(see previous paragraph), two AGN’s, and
Deficiency
15 galaxies show enhanced Balmer absorp-
tion lines but no emission, similar to PSBs Although gas is needed for star forma-
observed in z∼0.3 clusters. Most of the PSBs tion it is not obvious that only the atomic
in Coma are early type objects (mostly S0’s). gas content is related to star formation ac-
10