当前位置：文档库 › GMRT HI observations of the Eridanus group of galaxies

GMRT HI observations of the Eridanus group of galaxies

a r

s t

r o

J.Astrophys.Astr.(xxxx)xx ,000–000GMRT H i observations of the Eridanus group of galaxies A.Omar ??&K.S.Dwarakanath ?Raman Research Institute,Sadashivanagar,Bangalore 560080,India Received xxx;accepted xxx Abstract.The GMRT H i 21cm-line observations of galaxies in the Eri-danus group are presented.The Eridanus group,at a distance of ～23Mpc,is a loose group of ～200galaxies.The group ex-tends more than 10Mpc in projection.The velocity dispersion of the galaxies in the group is ～240km s ?1.The galaxies are clustered into di?erent sub-groups.The overall population mix of the group is 30%(E+S0)and 70%(Sp+Irr).The observa-tions of 57Eridanus galaxies were carried out with the GMRT for ～200hour.H i emission was detected from 31galaxies.The channel rms of ～1mJy beam ?1was achieved for most of the image-cubes made with 4hour of data.The correspond-ing H i column density sensitivity (3σ)is ～1×1020cm ?2for a velocity-width of ～13.4km s ?1.The 3σdetection limit of H i mass is ～1.2×107M ⊙for a line-width of 50km s ?1.Total H i images,H i velocity ?elds,global H i line pro?les,H i mass surface densities,H i disk parameters and H i rotation curves are presented.The velocity ?elds are analysed separately for the approaching and the receding sides of the galaxies.This data

will be used to study the H i and the radio continuum proper-ties,the Tully-Fisher relations,the dark matter halos,and the kinematical and H i lopsidedness in galaxies.

Key words :galaxies:groups,clusters –individual:Eridanus –radio lines:H i 21cm-line

2Omar&Dwarakanath

1.Motivation

Several redshift surveys carried out over the past several decades indicate that galaxies are distributed inhomogeneously in the local Universe.The regions of highest galaxy densities are superclusters and clusters.However, the majority of galaxies in the local Universe are found in less dense regions called groups.According to theories of hierarchical structure formation, clusters are built via mergers of groups.Clusters di?er from groups in sev-eral aspects.A remarkable di?erence is observed in the morphological mix and H i content of the galaxies.Clusters have an enhanced population of both the early type(S0and E)galaxies and the H i depleted spirals(Curtis 1918,Hubble&Humason1931,Davies&Lewis1973,Giovanelli&Haynes 1985,Warmels1988,Cayatte et al.1990,Bravo-Alfaro et al.2000)while groups are populated mainly by H i rich spirals.Dressler(1980)noticed a tight correlation between the galaxy morphology and the local projected galaxy density.This correlation,known as the density-morphology relation, is observed to be valid over more than?ve orders of magnitude in the pro-jected galaxy density(Postman&Geller1984).The origin of the enhanced population of E+S0’s in high galaxy density regions has been the subject of much debate.There are two hypotheses for the formation of S0’s,one is “Nature”where it is believed that early types were formed as such,and the other is“Nurture”according to which S0’s are transformed spirals.Some of the recent observations indicate that the clusters at intermediate redshifts (z～0.1?0.3)tend to have a higher fraction of S0’s at the expense of spirals (e.g.,Poggianti et al.1999,Dressler et al.1997,Fasano et al.2000).These observations support the“Nurture”scenario.

Several gas-removal mechanisms,viz.,ram-pressure stripping(Gunn& Gott1972),thermal conduction(Cowie&Songaila1977),viscous stripping (Nulsen1982),harassment(Moore et al.1998),starvation etc.have been proposed to explain the H i de?ciency in cluster spirals.Some of these pro-cesses are also believed to be driving the transformation of spirals to S0’s. Each of these processes has been predicted to remove H i mass of the or-der of the typical H i mass of a galaxy.These processes,however,are quite sensitive to several parameters like the density of the intra-cluster medium (ICM),the radial velocity vector of the galaxy,the magnetic?eld in the ICM,the gas-reservoir in the galaxy halo,etc..Some of these parameters have poor estimations which make the e?cacy of these processes doubtful. It has been argued that no single gas-removal mechanism can explain the global H i de?ciency in cluster spirals(e.g,Magri et al.1988).The ex-act physical mechanism(s)responsible for H i depletion in cluster spirals, therefore,remains uncertain.These di?culties have led one to speculate that cluster galaxies were perhaps H i de?cient even before they fell into the cluster.Such a scenario can be explored by studying groups of galaxies.

Several groups have been previously imaged in H i,e.g.,the Hickson

Eridanus group:H i observations3 Compact Groups(HCGs;Verdes-Montenegro et al.2001)and the the Ursa-Major group(Verheijen&Sancisi2001).The galaxy densities in HCGs are comparable to that in galaxy clusters,although HCGs have far less number of galaxies compared to that in clusters.The galaxies in some of the HCGs were found to be signi?cantly H i depleted.HCGs also tend to have a signi?cant population of early type galaxies.The Ursa-Major group,which has only a few S0’s and no ellipticals,showed no signi?cant H i de?ciency. The environment in the Ursa-Major group is similar to that in?eld.Here, we present an H i survey of the Eridanus group of galaxies with the recently completed Giant Meterwave Radio Telescope(GMRT).The Eridanus group is believed to be at an evolutionary stage intermediate to that of?eld and a cluster.The Eridanus group has a signi?cant population of early type galaxies.The sub-grouping of galaxies in the group is quite prominent.The Eridanus group also has weak di?use x-ray emission centered around some of the brightest galaxies in the sub-groups.On a broader perspective,the properties of the Eridanus group are between that of a loose group like the Ursa-Major and a cluster like Fornax or Virgo.The main aim of this survey is to identify the galaxy evolution processes active in an environment intermediate between that of a cluster and?eld.

The GMRT observations provided both the H i and the radio continuum (ν～1.4GHz)data.The kinematical information of galaxies has also been obtained using the H i data.This survey has capabilities to carry out sev-eral other studies.Some of the studies proposed to be carried out are the following-

?H i content of galaxies in the Eridanus group.

?H i morphologies of galaxies in the group.

?Tully-Fisher relations.

?Radio–Far-infrared correlation.

?Rotation curves and dark matter halos.

?Kinematical and H i lopsidedness.

In the present paper,the GMRT observations and the data analyses are described.We also investigate correlations between H i and optical proper-ties in this paper.The paper is arranged in the following order.The next section describes the properties of the Eridanus group.Sect.3contains de-tails of the GMRT observations.The analyses of the H i images are described in Sect.4.Some of the H i properties of the Eridanus galaxies are discussed in Sect.5.The results are presented in the tables in App.A.The H i atlas consists of the H i images,the H i velocity?elds,the global H i pro?les,the H i surface densities,the H i rotation curves,and the kinematical parameters of the H i disks.The H i atlases are given in App.B.

4Omar&Dwarakanath

2.The Eridanus group

2.1Introduction

The concentration of galaxies in the Eridanus region is known for many decades(Baker1933,1936).The complex morphology of this region was pointed out by de Vaucouleurs(1975).The Eridanus group was identi-?ed as a moderate size cluster in a large scale?lamentary structure near cz～1500km s?1in the Southern Sky Redshift Survey(SSRS;da Costa et al.1988).This?lamentary structure,which is the most prominent in the southern sky,extends for more than20Mpc in projection.The Fornax cluster and the Dorado group of galaxies are also part of this?lamentary structure.The dynamical parameters of a few galaxies in the Eridanus group were?rst published by Rood et al.(1970).With the increased number of identi?cations in this region by Sandage&Tammann(1975)and Welch et al. (1975),the latter authors speculated a dynamical connection between the Fornax cluster and the Eridanus https://www.wendangku.net/doc/75165796.html,ing the data from the Southern Galactic Cap sample(SGC;Pellegrini et al.1990),Willmer et al.(1989) grouped the galaxies in the Eridanus region into di?erent sub-groups and studied their dynamics.They concluded that each sub-group is a bound structure and possibly the entire group is also gravitationally bound with a dynamical mass greater than1013M⊙.They further pointed out that the Fornax and the Eridanus together constitute a bound system.The Eri-danus group is a dynamically young system with a velocity dispersion of ～240km s?1,which is lower compared to that(～1000km s?1)seen in clusters like the Coma.The distance to the group is estimated as23±2Mpc based on the surface brightness?uctuation measurements(Tonry et al.1997, Jensen et al.1998,Tonry et al.2001).All identi?ed members in the group are in the Helio-centric velocity range of～1000?2200km s?1,except NGC1400(S0),which has a velocity of～558km s?1.However,NGC1400 is predicted to be at a similar distance as that of the other members of the group based on the surface brightness?uctuation measurements.

2.2Group structure,membership,and morphological mix

The velocity-cone diagrams are plotted in Fig.1.The plots are in the Super-galactic coordinates.The velocities obtained from the NASA Extra-galactic Database(NED)are Heliocentric,and follow the optical de?nition.The clus-tering of galaxies near l=283?and b=?43?is the Eridanus group.Most of the galaxies are concentrated in the velocity range cz=1000?2200km s?1. The group appears to be loose and irregular.The clustering of galaxies near the apex(cz=0)is the local group.In Fig.2,the positions of galax-ies within the velocity range500?2500km s?1are plotted.There are181 galaxies in this plot,60early types(E+S0)and121late types(Sp+Irr).The approximate boundaries of three main sub-groups identi?ed by Willmer et

Eridanus group :H i observations

5500100015002000

2500

3000V

l o c i t y (k m /s )-60

-50-40

-30-20

Supergalactic latitude 500

10001500200025003000

V e l o c i t y (k m /s ) 270 280 290 300

Supergalactic longitude Figure 1.Positions of galaxies in the velocity-cone diagrams.The velocities are Heliocentric from the NASA Extra-galactic Database (NED).The circles mark the Eridanus group.

6Omar&Dwarakanath

Figure 2.Galaxies in the Eridanus group.The radius of the dashed circle is ～2Mpc within which galaxies were observed with the GMRT.A few sub-groups are marked with their approximate boundaries as identi?ed by Willmer et al.(1989). al.(1989)are marked in this?gure.The sub-clustering of galaxies can be seen in this?gure.

It can be seen that most of the early type galaxies are in the region in-side the circle marked in Fig.2.The sub-clustering is also prominent in the inner region.In the outer regions,population is dominated by spirals.The morphological mix is appreciably di?erent in each sub-group.The three sub-groups namely NGC1407,NGC1332,and NGC1395have their brightest members as an elliptical or an S0.The NGC1407sub-group is the richest in the early types,most of them being S0s.The population of(E+S0)and (Sp+Irr)in the NGC1407sub-group is70%&30%respectively,while that in most of the other sub-groups is～40%&60%respectively.The overall population mix of the Eridanus group is～30%(E+S0)&70%(Sp+Irr). The velocity-histograms of the early types and the late types are plotted in

and

Fig.3.There is no appreciable di?erence in the velocity range over which the early types and the late types are distributed.However,it can be seen from the upper panel of Fig.3that the distribution of early-type galaxies is approximately a Gaussian whereas the late-type galaxies are rather uni-formly distributed.Upon inspection of the locations of late-type galaxies

at lower(1000-1300km s?1)and higher(1800-2100km s?1)velocity ends

8Omar&Dwarakanath

from the mean,it appears that the galaxies having higher velocities are more uniformly distributed in the sky compared to those at lower velocities.The lower velocity galaxies are largely con?ned within the circle drawn in Fig.2. Further,it is interesting to note that the population mix of the NGC1407 sub-group is similar to that seen in evolved clusters like the Coma,whereas the velocity dispersion(～250km s?1)of the NGC1407sub-group is much smaller than that of Coma(～1000km s?1).Some other groups like Leo I, NGC3607,and NGC5846,which have lower velocity dispersions compared to that in clusters are also populated mainly by early type galaxies.

2.3X-ray properties

The optically bright early type galaxies NGC1400,NGC1407,NGC1395, and NGC1332are known X-ray sources in this group.Trinchieri et al. (2000)reported the presence of di?use X-ray emission around NGC1407. The processed and calibrated X-ray images(0.1keV-2.0keV)centered at NGC1407,NGC1395,and NGC1332were obtained from the ROSAT PSPC(Roentgen Satellite Position Sensitive Proportional Counter)archival data.Each?eld was observed for～6hour using the ROSAT PSPC instru-ments.The soft X-ray images shown in Fig.4were convolved with a circular Gaussian beam of FWHM90′′to enhance the di?use emission.Apart from the X-ray emission associated with NGC1400,NGC1407and a few other unresolved sources in the?eld,di?use emission centered at NGC1407and NGC1395can be seen in Fig.4.The extent of the di?use emission is～30′(～200kpc)around NGC1407and～20′(～135Kpc)around NGC1395. No di?use emission is seen around NGC1332(not shown here).Using the PIMMS(Portable,Interactive Multi-Mission Simulator;Mukai1993) tool,the di?use emission was modeled as thermal free-free emission from a Raymond-Smith plasma of energy～1.0keV(T～107K)and metallicity 0.2–solar.The choice of the temperature and the metallicity is in accordance with typical values found in X-ray groups(Mulchaey2000).

The total X-ray luminosity of the di?use emission in the energy range 0.1?2.0keV is1.6×1041erg s?1for the NGC1407sub-group and～6.8×1040erg s?1for the NGC1395sub-group.The intra-group medium density is estimated as～2.0×10?4cm?3in the X-ray emitting region.The X-ray luminosity of the Eridanus group is about2-3orders of magnitude lower compared to that of the clusters like Coma and Virgo.The estimated intra-group medium density is about an order of magnitude lower than that observed in virialised clusters like Coma.

2.4Comparison of the Eridanus group with other groups and clusters The properties of the Eridanus group are compared with the Virgo and the Fornax cluster,and the Ursa-Major group in Tab.1.All these systems are

Eridanus group:H i observations 9

NGC 1400

NGC 1407

30 kpc

NGC 1395

30 kpc

Figure4.Contours of X-ray emission around NGC1407and NGC1395overlaid upon the optical images from the DSS.The X-ray images are retrieved from the ROSAT PSPC archived data and smoothed with a circular Gaussian beam of90”.

10Omar&Dwarakanath

https://www.wendangku.net/doc/75165796.html,parison of four nearby galaxy groups and clusters

Properties Virgo a Fornax b Eridanus c Ursa-Major d Notes-(a):Inner6?region,(b):Inner2?.4region,(c):Inner9?region,(d):Inner15?region.Ref-erences-(1)Federspiel et al.(1998),(2)Ferguson(1989),(3)Binggeli et al.(1987),(4)Mushotzky &Smith(1980),(5)Mould et al.(2000),(6)Richter&Sadler(1985),(7)Paolillo et al.(2001),(8) Omar(2004),(9)Sakai et al.(2000),(10)Tully et al.(1996)

at comparable distances.Both the Fornax cluster and the Eridanus group

belong to the?lamentary structure described by da Costa et al.(1988).

The Ursa-Major group is a loose group of galaxies(Tully et al.1996).All

the four systems have quite di?erent properties.The Fornax cluster having

the highest galaxy density has the lowest spiral fraction,consistent with the density-morphology relation.The Eridanus group is intermediate between

the Virgo cluster and the Ursa-Major group in terms of its velocity disper-

sion,its x-ray luminosity,and its number of early-type galaxies.The mean projected galaxy density in Eridanus is intermediate between that in Ursa-

Major and in Virgo.The galaxies in Virgo are H i de?cient.The Ursa-Major

group has normal H i content.The X-ray luminosity of the Eridanus group

is at the lower end of the X-ray luminosities observed in groups(Mulchaey 2000).The velocity dispersion of the galaxies in the Eridanus group is inter-mediate between that in Fornax and in Ursa-Major.From this comparison,

it appears that the Eridanus group forms a system which is intermediate between a loose group(Ursa-Major)and a rich cluster(Virgo and Fornax).

3.Observations and data reduction

The present GMRT H i observations o?er several advantages over studies

carried out in the past using single dish telescopes.The GMRT is an in-terferometric array of thirty45-m diameter fully steerable parabolic dishes.

A description of the GMRT is given by Swarup et al.(1990).The GMRT

is located at a site(longitude=74?.05E,latitude=19?.092N,height

Eridanus group:H i observations11～650m above MSL)about80km north of Pune,India.The con?guration of the GMRT is optimized to meet the requirements of high angular reso-lution and of being able to image extended emission.This optimization is achieved through a hybrid con?guration of the antennas.Fourteen of the thirty dishes are located more or less randomly in a compact central array within an area of about1×1square kilometer,and the remaining sixteen dishes are spread out along the3arms of an approximately Y shaped con?g-uration over a larger region.The longest separation of antennas is～25km, and the shortest separation is～100m.The GMRT is expected to be sen-sitive to structures on the scales of2′′?7′at a wavelength of21cm.The angular sizes of the galaxies in the Eridanus group are in the range1′?5′implying that the data should be sensitive to image radio emission(H i and continuum)over the full extents of galaxies.The FWHM of the primary beam of a GMRT antenna is～24′at1.4GHz.

3.1Sample of galaxies

The selection of galaxies for the H i observations were made keeping in mind the broad perspective of the work.The galaxies were not selected based on their H i contents or their optical luminosities.Both early type and late type galaxies were included in the sample.The galaxies were selected from the inner4Mpc region of the group where galaxy density is higher and most of the S0’s are found.A follow up R-band photometric observations were also carried out with the1-m optical re?ector at the Aryabhatta Research Institute of Observational Sciences(ARIES;formerly State Observatory), Nainital.The optical data analysis is presented in Omar(2004).

Since the present study was carried out with a limited telescope time of ～200hour,the pointing centres of the observations were adjusted in a way to include two or more galaxies within the FWHM of the primary beam.Unfor-tunately,one complete run of observations on16galaxies(during November 2001),mostly early types,was badly a?ected due to ionospheric scintilla-tions,perhaps related to the intense solar activities during that year.The data collected during this period could not be used to obtain images.Five galaxies from these lost observations were re-observed later in2002.The science quality data were obtained for a total of46galaxies.In Tabs2& 3,the complete observed sample of57galaxies is listed with some of their previously known optical and radio properties.

3.2Observational parameters

The Eridanus group can be observed with the GMRT for～8hour in a given day.Often,two galaxies were observed in each day.The observing strategy was optimized to get uniform distribution of visibilities.Two galaxies were observed alternately for15?20minute each followed by5?7minute of

12Omar &Dwarakanath

D E C L I N A T I O N (J 2000)RIGHT ASCENSION (J2000)03 5550454035302520-15

-16-17

-18-19-20-21-22

-23-24NGC 1407NGC 1395

NGC 1332

NGC 1400123

47658910

1112

1314

15161718

1920

2122232425

26272829

S0E Sp ?Figure 5.The ?elds observed with the GMRT.The bigger circles correspond to the FWHM of the GMRT primary beam (～24′)at 1.4GHz.

observations of secondary calibrators.This cycle was repeated and a total of 3-4hour of observing time was accumulated on each galaxy.Most of the observations were carried out using an 8MHz bandwidth over 128channels,which gives a velocity resolution of ～13.4km s ?1.The observations were carried out for longer duration (～8hour)for some of the early type galax-ies,and with smaller bandwidths (2-4MHz)for smaller inclination galaxies to get su?cient velocity resolution.A total of ～200hour of the GMRT observations were carried out spread over a period of 2years (2000–2002).

Eridanus group:H i observations13 The data obtained during November,2001which were corrupted due to scintillations were discarded.The observing parameters are listed in tab.4.

The VLA calibrators0240-231and0409-179were used as the secondary calibrators.0240-231is classi?ed as“un-resolved”for all the four VLA con-?gurations with a20-cm?ux density of6.3Jy.0409-179is resolved by baselines longer than10km with a20-cm?ux density of2.2Jy.0137+331 (3C48)and0542+498(3C147)were used as the primary calibrators.3C48 is resolved by baselines longer than8km with a20-cm?ux density of16.5Jy and3C147is resolved by baselines longer than10km with a20-cm?ux density of22.5Jy.3C48was observed in the beginning and3C147was ob-served at the end of each observing run for20-30minutes.The?ux densities of the primary calibrators were estimated at the observed frequencies using their known radio spectra from the VLA observations in the1999.2epoch.

3.3Data acquisition and reduction

The data(visibilities)were collected in the LTA(Long Time Accumulation) format,which is the native format for the GMRT data.The LTA data were converted to FITS(Flexible Image Transport System)format for subsequent processing.Visibilities were averaged over～16second.The data were monitored on line.The data were later?agged from the antennas having low gains,for time ranges where data were corrupt,and at lower elevations (usually below25?)where correlation drops signi?cantly(below50%in some cases).The?ux densities of the secondary calibrators were estimated based on the?ux densities of the primary calibrators.The visibilities on the target galaxies were calibrated by interpolating the complex gains determined using the secondary calibrators.Since the spectral responses of?lters are not?at, the initial calibration was carried out using the data averaged over four to six channels.The spectral response of the antennas were determined using both the secondary and the primary calibrators,and an averaged spectrum was used to correct the band shapes.The gains start declining signi?cantly after the110th channel.The?rst1?3channels are generally corrupted in the?lter response.Therefore,the data were used between channels3?115. An initial H i spectrum was generated using the AIPS(Astronomical Image Processing System)task POSSM at spatial frequencies below2kλin the direction of target galaxies.This range of spatial frequencies enables most of the H i signal to be detected in the H i spectrum.This spectrum is to identify channels with H i emission.

The continuum-data were generated by averaging the channels devoid of H i line emission.The continuum images were made using this channel averaged data and were used for self-calibration.Several iterations of phase self-calibrations were performed to improve the dynamic range of the im-ages.The?nal self-calibrated solutions were applied to the line-data.The continuum emission was subtracted from the line-data using the AIPS tasks

14Omar&Dwarakanath

UVSUB and UVLIN.The self-calibrated and continuum subtracted line-data were used to make the image cubes at di?erent resolutions by selecting ap-propriate(u,v)ranges.The image cubes were made at two resolutions-one with a resolution of～15′′(high resolution cube)using(u,v)data in the range0.2?20kλ,and another with a resolution of～50”(low resolution cube)using(u,v)data in the range0.2?5kλ.

4.Image analysis

The images were analysed using the GIPSY(Groningen Image Processing System)package developed by the Kapteyn Institute,the KARMA visual-ization tool(Gooch1996),and the AIPS package developed by the National Radio Astronomy Observatory.

Since the angular resolution varied by a few arc second in di?erent cubes, all high resolution cubes were convolved to a common resolution of20′′×20′′. In some cases,intermediate resolution cubes at25′′or30′′were also made. The channel images typically have an rms of1mJy beam?1.The3σcolumn density detection limit in the20′′images is1×1020cm?2.The cubes are sensitive(3σ)to detect a galaxy of H i mass1.2×107M⊙for an H i line-width of50km s?1.The image cubes were inspected visually to identify H i signals. The channel images are presented for all the H i detected galaxies elsewhere (Omar2004).An example of the channel images is shown in Fig.6.

4.1Total H i map and H i diameter

The zeroth and the?rst order moment maps were generated at both the low(50′′)and the high(20′′)angular resolutions.The moment zero map or the total H i image is obtained by summing the H i images in di?erent channels.The cubes were?rst blanked to separate the H i signals from noise before summing the channels.The blanking can be done in several ways. The total H i image depends on the blanking procedure(Rupen1999).One of the methods is to blank the pixels below a certain?ux density level.

A higher cuto?(e.g.,5σ)makes the total H i images patchy while a lower cuto?(e.g.,3σ)makes the images noisy.The low surface brightness nature of H i emission makes it di?cult to separate the low level signals from noise.A hybrid approach has been shown to be e?ective in overcoming this problem (Rupen1999).This approach involves masking the noise after smoothing the cube using the AIPS task MOMNT.The moment maps are still estimated using the un-smoothed cube in this approach.

The?ux density(mJy beam?1)is converted to H i column density using the following relation(eqn.3.38,Spitzer1978):

Eridanus group :H i observations

15-21 27

2017.1 Km/s 2003.8 Km/s 1990.4 Km/s 1977.0 Km/s -21 27

1963.7 Km/s 1950.3 Km/s 1936.9 Km/s 1923.6 Km/s -21 27

1910.2 Km/s 1896.9 Km/s 1883.5 Km/s 1870.2 Km/s D E C L I N A T I O N (J 2000)-21 27282930

1856.8 Km/s 1843.4 Km/s 1830.1 Km/s 1816.7 Km/s -21 27

1803.4 Km/s 1790.0 Km/s 1776.7 Km/s 1763.3 Km/s 03 33 50454035-21 27

1750.0 Km/s 1736.7 Km/s RIGHT ASCENSION (J2000)03 33 504540351723.3 Km/s 1710.0 Km/s

Figure 6.H i emission from IC 1953.The ‘+’sign marks the optical centre of the galaxy.The rms/channel is 1.4mJy beam ?1.The contours are at 2.5,3.75,5,7.5,and 10times the rms.The images are convolved with a circular Gaussian beam of 20′′×20′′.

16Omar &

Dwarakanath

N 1371

E 482?005

IC 1952

N 1347

E 549?035N 1481

N 1482

E 549?002

M ?03?10?041

N 1309

N 1345

U 077N 1390E 548?072 E 548?065N 1359

S 0321.2?1929

U 068E 548?021N 1325

E 482?013

N 1385

E 482?035

N 1415

N 1414N 1422 E 548?036IC 1953E 548?049IC 1962

E 549?018Figure 7.A contour-image collage of the H i detected galaxies in the Eridanus group.Only one contour is plotted to indicate the extent of each galaxy at N H i =1020cm ?2.The individual galaxies are magni?ed ten times.To avoid overlap,some galaxies are slightly displaced from their actual positions.A bar at the upper right hand corner indicates a scale of 20kpc for the enlarged sizes of the galaxies.Otherwise,1?corresponds to ～400kpc.

N H i (α,δ)=1.1×1021cm ?2

Eridanus group :H i observations

020 kpc ?2

Figure 8.A collage of the H i detected galaxies in the Eridanus group.The color key indicates the H i column density.

The diameters of the H i disks were estimated from the high resolution total H i images at a ?xed face-on H i surface density of 1M ⊙pc ?2.Due to projection e?ects,the sensitivities to the face-on H i surface densities were not uniform.Therefore,in some cases the H i diameters were extrapolated to the face-on H i surface density level of 1M ⊙pc ?2.

4.2H i velocity ?eld

The conventional way of deriving the velocity ?eld is to compute the intensity-weighted ?rst order moment of the H i images at di?erent velocities.There

18Omar&Dwarakanath

is an alternative to obtain velocity?elds by?tting the H i pro?les at ev-ery pixel with a Gaussian.These pro?les are usually asymmetric depending upon the kinematics of H i along the line-of-sight,and also due to the beam smearing caused by the?nite size of the synthesized beam.The broadening and the asymmetry will depend upon the H i distribution in the galaxy.The e?ect of beam smearing will be more pronounced in edge-on systems.As a result of the asymmetry and the broadening in the H i pro?le,a single Gaussian component will not give an accurate result.Unfortunately,multi-component Gaussian?t could not be carried out as the signal to noise ratio of the detections were not su?cient.

Rupen(1999)has brie?y discussed the merits and drawbacks of these two procedures.In the present analysis,the?rst order moment maps were found to be generally noisier than the velocity?eld maps obtained by Gaussian?ts. This may be because the Gaussian?ts were not as sensitive to the outliers as moment map were.Therefore,in the present analysis,Gaussian?ts were used to construct the velocity?eld maps.It should be noted that both the procedures to obtain the velocity?eld will underestimate rotation velocities at locations of steep velocity gradients.The?at part of the rotation curve, however,remains una?ected.

4.3Global H i pro?les

The integrated H i?ux density as a function of velocity is the global H i pro-?le as would have been obtained from a single dish observation.The low resolution(50′′)cubes were used to obtain the global pro?les as these cubes are most sensitive to the di?use emission.The lower resolution zeroth order moment map was used to mark the region over which the?ux density was estimated in the channel images.The H i mass is obtained by using the following relation:

M H i(M⊙)=2.36×105D2δv N chan

j=1

S j(2)

where D is the distance in Mpc,S j is the integrated?ux in Jy in the spectral

channel j of velocity widthδv in km s?1.The distance is taken as23Mpc.

In Fig.9,the integrated H i?ux(δv N chan j=1S j)from the GMRT is compared with that from the HIPASS data and from other single dish data.Some of

the galaxies with higher values of the integrated H i?ux in the single dish data show signi?cantly less?ux in the GMRT.The H i disk sizes of these galaxies are among the largest(>6′)in our sample.We believe that the loss of?ux for the large galaxies in the GMRT images is due to inadequate sampling of shorter(u,v)spacings in the GMRT.

Eridanus group:H i observations19

Figure9.A comparison of the integrated H i?ux densities of Eridanus galaxies from the GMRT with those from the single dish data published elsewhere and from the HIPASS.Most of the ratios are within±25%of unity(indicated by the dotted lines).

4.4H i line-width

The H i global pro?les often peak at the two extreme ends of the rotation velocities of galaxies.The detailed shape of an H i pro?le depends on the rotation curve,the inclination and the H i distribution in the galaxy.The H i line-widths are broadened from their true values due to random motions in the H i gas and due to the?nite spectral resolution.The H i line-width is a crucial parameter for studying the Tully-Fisher(TF)relation.Most of the TF studies still use the H i line-widths obtained from single dish observations because of the simplicity of such observations.The synthesis data of the Eridanus group of galaxies provide an opportunity to compare the corrected H i line-widths obtained from the single dish H i pro?les with those obtained from the H i rotation curves.

For those cases in which double-peaked H i pro?les are seen,H i line-widths were estimated at20%(W20)and at50%(W50)levels of the peak intensities at the two ends of the H i pro?les.The locations of the two peak intensities were estimated separately using Gaussian?ts to the pro-

20Omar&Dwarakanath

?les.Bottinelli et al.(1990)derived an empirical relation to correct for the instrumental broadening.They convolved the H i pro?le progressively with coarser velocity resolutions for a model galaxy,and determined the broad-ening.A linear relationship between the channel resolution and the instru-mental broadening was suggested.The broadening correction is estimated asδW=0.55×δV i for W20andδW=0.13×δV i for W50for an instrumental resolution ofδV i.For the current observations,δV i=13km s?1,implying that the corrections are～7km s?1for W20and～2km s?1for W50.

A linear summation of the rotation velocity and the random velocity is appropriate to estimate the observed width for the cases where the intrinsic width is almost boxy(i.e.,in fast rotating galaxies).However,a summation in quadrature will be required for the slow rotating(e.g.,dwarf)galaxies where the solid body rotation together with the radial distribution of the H i gas will lead to an almost Gaussian pro?le.A composite relation for all galaxies was given by Tully&Fouque(1985).According to their relation, the width due to the rotation motion W R,the width due to random and turbulent motions W t,and the observed width W l are related by:

W2l=W2R,l?W2t,l 1?2e?(W l/W c,l)2 +2W l W t,l 1?e?(W l/W c,l)2 (3) where the subscript l refers to the level(20%,or50%)at which the widths are estimated.The W t,l is estimated as2k lσfrom the velocity dispersion of the H i gas(σ)due to random and turbulent motions.The constant factor k l is1.80at20%level and1.18for the50%level for a Gaussian pro?le. The value ofσis taken as6km s?1.W c,l is a parameter which de?nes the transition region from linear to quadratic sum.The eqn.3does a linear subtraction if W l>W c,l and a quadratic subtraction if W l

The corrected H i widths are compared with the?at rotation velocities of the galaxies in Fig.10.The mean value of the di?erence(W R,50?2V flat) is～6.5km s?1.

4.5Rotation curves

The rotation curves were derived using the tilted ring model(cf.Begeman 1989).The GIPSY task ROTCUR was used.The basic methodology of this model is the following.The model assumes the gas to be in circular orbits.The position angle and the inclination of the H i disk are allowed to vary with radius.The?tting procedure generally involves estimation of5 unknowns,viz.,the dynamical centre(X,Y),the systemic velocity(V sys), the position angle(PA)of the major axis,the inclination angle(Incl),the circular rotation velocity(V rot),and optionally the expansion velocity V exp.