NGC1600 - Cluster or Field Elliptical?
Rodney M. Smith, Vicent J. Martinez, Alberto Fernandez-Soto, Fernando J. Ballesteros, Amelia Ortiz-Gil
aa r X i v : . [ a s t r o - ph ] F e b The Astrophysical Journal, 679: ? - ?, 2008 June 1 (scheduled)
Preprint typeset using L A TEX style emulateapj v. 08/22/09
NGC1600 - CLUSTER OR FIELD ELLIPTICAL?
Rodney M. Smith , Vicent J. Mart´ınez , Alberto Fern´andez-Soto , Fernando J. Ballesteros , andAmelia Ortiz-Gil The Astrophysical Journal, 679: ? - ?, 2008 June 1 (scheduled)
ABSTRACTA study of the galaxy distribution in the field of the elliptical galaxy NGC1600 has been undertaken.Although this galaxy is often classified as a member of a loose group, all the neighbouring galaxies aremuch fainter and could be taken as satellites of NGC1600. The number density profile of galaxies inthe field of this galaxy shows a decline with radius, with evidence of a background at approximately1.3 Mpc. The density and number density profile are consistent with that found for other isolatedearly-type galaxies. NGC1600 appears as an extended source in X-rays, and the center of the X-rayemission seems not to coincide with the center of the galaxy. The velocity distribution of neighbouringgalaxies has been measured from optical spectroscopic observations and shows that the mean radialvelocity is approximately 85 km s − less than that of NGC1600, indicating that the centre of masscould lie outside the galaxy. The velocity dispersion of the ‘group’ is estimated at 429 ±
57 km s − .The inferred mass of the system is therefore of the order of 10 M ⊙ , a value that corresponds to alarge group. NGC1600 therefore shares some similarities, but is not identical to, the ‘fossil clusters’detected in X-ray surveys. Implications of this result for studies of isolated early-type galaxies arebriefly discussed. Subject headings: galaxies: elliptical and lenticular INTRODUCTION
The importance of isolated early-type galaxies in un-derstanding galaxy and cluster evolution has led to asurge in the available number of catalogues of such ob-jects (e.g Colbert et al. 2001, Smith et al. 2004, Redaet al. 2004, Denicolo et al. 2005, Buote 2005), with fewgalaxies in common between them. The variation arisesfrom the different selection criteria used in the catalogueconstruction. A common requirement is that there is nogalaxy with a similar redshift within a certain distance,typically of the order of 1 Mpc. However, the incomplete-ness of current galaxy catalogues, and in particular thelack of complete redshift information, usually requires avisual inspection of the field as final confirmation thatthe galaxies are isolated. The catalogue of Smith et al.(2004, hereafter SMG) only used redshift information forthe central galaxy and discounted any galaxies which hada bright galaxy nearby in projection. It thus contains thestrictest isolation criteria and all galaxies in the sampleare the most dominant member in the field by at least 2.2magnitudes within 500 kpc and 0.7 magnitudes within1 Mpc. Again, visual inspection confirmed that thesegalaxies were indeed isolated. The SMG catalogue willmiss several galaxies in the other samples as no redshiftinformation is applied. However, interpretation of the lo-cal faint galaxy environment is simplified by the lack ofother bright galaxies nearby which may have their owndwarf population.
Electronic address: [email protected] Dept. of Physics and Astronomy, Cardiff University, Cardiff,CF24 3AA, U.K. Observatori Astron`omic, Universitat de Val`encia, Apartat deCorreus 22085, E-46071 Val`encia, Spain Departament d’Astronomia i Astrof´ısica, Universitat deVal`encia, Apartat de Correus 22085, E-46071 Val`encia, Spain
NGC1600, an 11.9 B -magnitude ( K =8.1 ) E3-E4 gal-axy at a distance of approximately 64 Mpc (NED andusing H = 73 km s − Mpc − ) was included in the sam-ple of SMG due to the large magnitude difference be-tween it and other nearby galaxies, even though it is oftenquoted as being at the centre of a loose group containinga number of NGC-catalogued galaxies together with sev-eral other fainter galaxies. Denicolo et al. (2005) classifyit as a field elliptical rather than isolated. The brightestgalaxy within a projected distance of 1 Mpc is NGC1594,a 13.9 B -magnitude ( K =10.2) galaxy at least 800 kpcaway. The two closest NGC galaxies, NGC1601 andNGC1603 at separations of 1.6 and 2.6 arcminutes (30kpc and 48 kpc) respectively, are both almost three mag-nitudes fainter than NGC1600. Thus, although there areseveral neighbouring galaxies in the field of NGC1600,it is by far the dominant galaxy in the group. There isa poor group of galaxies (Zwicky cluster 0430.8-0424B,Zwicky et al. 1961-1968) at a distance of 53 arcmin to theNorth of NGC1600, with a mean velocity of 5030 km s − (Baiesi-Pillastrini et al. 1984), a velocity similar to thatof NGC1600 (4715 km s − ). The brightest galaxy in thisgroup still satisfies the isolation criteria of SMG, i.e. itis more than 2.2 magnitudes fainter than NGC1600.The galaxy has been the subject of several previousstudies. In a photometric study, Matthias and Ger-hard (1999) found the galaxy to have boxy isophotes,whilst kinematic studies (Bender et al. 1994, Faber etal. 1989) found that there was little rotation in the stel- All near-infrared magnitudes in this paper have been obtainedfrom the 2MASS catalogues. The NASA/IPAC Extragalactic Database (NED) is operatedby the Jet Propulsion Laboratory, California Institute of Tech-nology, under contract with the National Aeronautics and SpaceAdministration.
Smith, R.M. et al.lar component, with a maximum of 30 km s − , whilstthe central velocity dispersion is typical of larger ellip-tical galaxies (321 km s − ). There is evidence of pastand possible ongoing star formation, with the presenceof H α regions (Trinchieri and di Serego Alighieri 1991)and dust (Ferrari et al. 1999). Thus it is highly likelythat NGC1600 is a merger remnant (Matthias and Ger-hard 1999) although, with broadband colours typical ofelliptical galaxies (NED), then this merger is likely tohave occurred a long time ago. Estimates of the ageof this galaxy hence range from 4.6 Gyr up to 8.8 Gyr(Trager et al. 2000, Terlevich and Forbes 2002).Theseresults are in agreement with the study of a sample ofisolated early-type galaxies by Reda et al. (2004), whofound that several galaxies had boxy isophotes, evidenceof dust and past merger activity but had remained rela-tively dormant for several Gigayears.NGC1600 is also a weak X-ray source (Sivakoff et al.2004), showing extended emission out to at least 100 arc-sec (corresponding to a physical radius of 30 kpc , witha possible central component associated with NGC1600and an outer emission region associated with the group.The extended outer emission is centred to the North-Eastof the central galaxy, suggesting that the centre of thepotential is slightly offset from NGC1600. The presenceof extended material around the galaxy is also suggestedfrom the tailed X-ray structure around the nearby galaxyNGC1603, indicating an effect of ram-pressure stripping.The large magnitude difference between NGC1600 andthe other galaxies in its neighbourhood, together with theextended X-ray emission, would suggest that NGC1600may be a fossil group.Whereas the radius of about 200 arcsec, or 60 kpc, to-tal soft X-ray emission of 2 × erg s − and absolutemagnitude of M R = − . M ⊙ (e.g. Parker et al., 2005)whilst spiral galaxies have masses an order of magnitudesmaller, even allowing for any dark matter componentat large radii (e.g. Zaritsky et al. 1997). The radii ofgroups is also an order of magnitude greater than thatof the individual galaxies themselves (e.g. Karachentsev2005). The mass and extent of elliptical galaxies is atpresent very uncertain, with X-ray evidence that someare surrounded by dark matter haloes (e.g. Fukazawa et al. 2006) whilst optical evidence may suggest not,with masses similar to that of spiral galaxies (e.g. Ro-manowsky et al. 2003). However, the large differencebetween the likely mass and extent of galaxy groups com-pared to isolated galaxies should enable us to distinguishbetween the group-member and the isolated elliptical hy-potheses for NGC1600 through a dynamical study of theother neighbouring galaxies.The presence of X-ray emission around NGC1600can provide an estimate of the total mass of the gal-axy, assuming hydrostatic equilibrium within the gas.Fukazawa et al. (2006), analyzing Chandra observationsof a number of elliptical galaxies, determined a mass-to-light ratio for NGC1600, at the effective radius of 14.5kpc, of 10.58 M ⊙ /L ⊙ , corresponding to a total mass ofapproximately 10 M ⊙ . This is higher than that of manygalaxies in their sample.An investigation of the spatial distribution of galax-ies around NGC1600 may also help in distinguishing be-tween the isolated and group hypotheses. SMG found aweak excess of galaxies around isolated elliptical galax-ies out to at least 500 kpc, with an exponential slope of − . ± .
2. This is similar to that found for poor groupsand around individual galaxies but is less steep than thatfound for clusters (e.g. Hansen et al. 2005). The numberdensity of bright galaxies around isolated galaxies, how-ever, is much lower than that found around groups andclusters. OBSERVATIONS AND DATA REDUCTION
To obtain an estimate of the radial number density pro-file of galaxies around NGC1600 we have used the APMscans of UKST sky survey plates of a circular area ofradius 1.5 degrees (corresponding to a scale of 1.65 Mpcat the distance of NGC1600) surrounding this galaxy.Only those objects brighter than 20th magnitude in B and classified as galaxies in both B and R were selected.NGC1600 lies within 1 degree of the edge of plate 764in the survey and thus the adjoining plate, 765, mustbe matched in both position and magnitude. Using theTOPCAT package in the Starlink suite of astronomicaldata reduction packages we have used the matching ob-jects detected by the APM in the overlap region of plate764 and 765 to determine the accuracy of matching bothin magnitude and position. The average magnitude dif-ference between galaxies that lie in the overlap regionthat are detected in both plates is less than 0.05 magni-tudes and thus no magnitude difference is assumed. Inaddition, the number of matched objects varies by lessthan 5% with varying the angular distance between 1and 10 arcsec for those objects which are classified asmatched. This indicates that the positional accuracy ofthe scans is less than 1 arcsec. However, due to severalerrors such as misclassification, vignetting and magni-tude errors for bright galaxies, inconsistencies will arisebetween the galaxy sample derived from each plate, eventhough the mean density is almost identical and the mag-nitudes and positions agree for duplicate objects. Theplates cannot therefore be merged as this leads to an in-crease in the number density of galaxies within the over-lap region. Thus, to overcome this problem and reducevignetting problems at the edge of the plate, half theoverlap region was taken from plate 764 and half fromplate 765.GC1600 3The APM scans, as in all automatic object detectionmethods, have difficulty with not only star-galaxy sepa-ration at faint magnitudes but also the correct selectionof objects within the haloes of bright stars and galax-ies. For example, diffraction spikes are quite often splitinto several individual objects and their non-circular na-ture often leads to such structure being classified as non-stellar. There are several bright stars with diffractionspikes on the UKST plates of the field around NGC1600.These are not in the immediate neighbourhood of thecentral galaxy. However, to ensure that the effect of mis-classification and extraneous objects was insignificant wehave overlain the positions of the objects detected by theAPM over the UKST image for a visual inspection of thedetected objects. No objects were found correspondingto diffraction spikes from bright stars. The star-galaxyseparation technique used by the APM is also not 100%accurate and is heavily dependent upon the magnitude ofthe objects, becoming highly significant beyond B = 20(e.g. Maddox et al. 1990). Thus although it is likely thatsome objects are mis-classified, by limiting the sample to B = 20 the percentage of misclassifications is small andthis is confirmed by a visual inspection of the detectedobjects. A study of the detected objects superimposed onthe UKST Sky Survey plates showed that several of thebright galaxies in the field of NGC1600 were not in thecatalogue derived from the APM scans. Many of theseobjects were classified as merged objects by the APM.To estimate the effect of these missed galaxies on thedetermination of the radial number density profile, theseobjects were added by eye, discounting faint objects thatthe eye has difficulty classifiying and that may also failthe magnitude cut-off. This subjective technique intro-duces more errors into the analysis but will lead to someestimate of the true errors in the APM number densities.For the dynamical study, all galaxies within a 25 ar-cminute radius of NGC1600 were selected from the APMcatalogue (e.g. Lewis and Irwin 1996) of the UKST sur-vey plates. Each object had to be detected as a galaxy inboth B and R to be selected and a further visual checkwas made to ensure the correct identification as a galaxy.A magnitude cut of R = 20 was made to facilitate thepossibility of getting a redshift. As mentioned above,several bright galaxies were not in the APM galaxy cat-alogue due to misclassification and these were added tothe sample by hand.The field was observed during the nights of 7th Jan-uary 2003 and 12th January 2005 using the AUT-OFIB2/WYFFOS multi-fibre spectrograph on the 4.2-m William Herschel telescope (WHT) on the island ofLa Palma. During the 2003 run the large, 2.7arcsec, fi-bres were used to maximise the signal from the extendedsources whereas in 2005 only the small, 1.6 arcsec, fibreswere available. Additionally, in 2003 the CCD used onlyenabled 120 fibres to be used whilst in 2005 the CCDhad been replaced by a two 2k x 2k EEV CCD mosaicenabling the number of fibres available for objects to beincreased to 150. The smaller CCD used in 2003 sufferedfrom contamination of neighbouring spectra when brightobjects were observed. As most of the galaxies observedhere were faint this was not a significant problem.The AF2 configure program was used to configure theobservations, with the brighter galaxies, and those galax-ies within the un-vignetted field of 20 arcminutes radius, preferentially selected for observation. With the limi-tations imposed by the instrument it was impossible toobserve all galaxies in the field during one run. At least4 stars in the field were selected for guiding through thefiducial bundles, and unused fibres were placed on areasof the background field for sky-subtraction. The datafrom the 2003 run were reduced and the galaxies forwhich no redshift was obtained, together with a selec-tion of previously unobserved galaxies, were selected forthe 2005 run. On both runs, a wavelength range of ap-proximately 3850˚A to 5450˚A was covered, with a pixelsize of 0.4˚A. Wavelength calibration was warranted byfrequent observations of a neon and helium arc. Seeingduring the 2003 run was typically about 1 arcsec and atotal of 5400 seconds integration time was obtained onthe field. In 2005 the seeing was 1.4 arcsec on averagewith 7200 seconds of integration on the field.Due to the arrangement of the fibres on the spectro-graph, the observed spectra are not uniformly placed onthe CCD but are arranged in rows of 3, with a shift of 60pixels in the dispersion direction between three consecu-tive fibres. This leads to complications in the data reduc-tion and so the data were reduced using the Observatory-supplied wyffosREDUC IRAF package. This packagebias-subtracts, extracts the fibres, and subtracts the skyin an automated fashion. However, to ensure the reduc-tion was satisfactory, each step was visually checked. Asthe sky-subtraction in this package is not ideal inspec-tions of the spectra were made to ensure satisfactory skysubtraction, with regions around possible sky emissionlines removed. Redshifts were obtained using the IRAFxcsao package, with cross-correlation against a range oftemplates. RESULTS AND DISCUSSION
Figure 1 shows the radial number density profile ofgalaxies centred on NGC1600 estimated from the APMscans of the UKST sky survey plates of the region. Thereis a decline in the number density with radius and astrong suggestion that the background population hasbeen reached at a radius of about 1.3 Mpc. The errorsdisplayed are calculated assuming Poissonian statisticsand are thus are likely to be an underestimate. Misclassi-fication and problems with measuring bright galaxies willalmost certainly increase the error estimates, whilst inthe inner 150 kpc the small number of detected galaxiesand the blanking factor due to the presence of NGC1600will lead to greater uncertainties. The difference betweenthe profile determined purely by the APM detections andthat with galaxies missed by the APM but selected byeye included lies within the errors.Figure 1 clearly shows a decrease in the number densityof galaxies with galactocentric distance. The backgroundgalaxy population density appears to be reached at agalactocentric distance of about 1.3 Mpc, for a numberdensity of approximately 150 galaxies per square Mpc.Fitting an exponential to this distribution gives a slope of − . ± .
3. Although slightly steeper than the − . ± . − . Fig. 1.—
Projected density of galaxies brighter than B = 20around NGC1600. The solid line includes galaxies selected by eye,whilst the dotted line is only those galaxies detected by the APM.At the redshift of NGC1600, B = 20 is equivalent to M B = − that only the faint dwarf galaxies, with M R & − . B magnitude of −
14 andtherefore we would expect to see the excess of neighbours.By increasing the limiting magnitude to R = 18, corre-sponding to an absolute magnitude of M R = −
16 thenumber densities decrease significantly and the error barstherefore increase. Figure 2 shows the resultant densityprofile for this brighter magnitude limit. Although thereis evidence of a small excess of galaxies at small galac-tocentric distances, and these objects are seen by visualinspection of the field, the error bars are large enoughsuch that a uniform distribution, as suggested by Redaet al. (2004), is not ruled out.The radial density profile thus indicates that there isa significant population of faint ( M B & −
15) galaxies inthe neighbourhood of NGC1600 similar to that found,in general, from a sample of isolated ellipticals by SMG.These results are in agreement with the results of Redaet al. (2004) who suggest that apparently isolated galax-ies do not have a large number of bright galaxies in theirneighbourhood but could be surrounded by fainter galax-ies. The total excess of galaxies within 500 kpc downto M B = − . Compared to the background reached at approximately 1.3Mpc, with a number density of approximately 150
B <
20 galaxiesper square Mpc, as seen in previous paragraphs.
Fig. 2.—
Projected density of galaxies brighter than R = 18around NGC1600.The solid line includes those galaxies selected byeye whilst the dotted line represents only those galaxies detectedby the APM. shifts are shown in Table 1. Previously measured red-shifts are also given, together with those galaxies listedin NED and HyperLeda with redshifts between 3700 and5700 km s − and within a projected distance of 57 ar-cminutes of NGC1600 (corresponding to a physical pro-jected distance of 1 Mpc). Note that this latter list in-cludes members of the galaxy group Zw0430.8-0424B cat-alogued by Zwicky et al. (1961-1968).With a total of 30 known galaxies within 1000 km s − of NGC1600, our sample indicates that a very significantpopulation of galaxies exists in the physical neighbour-hood of NGC1600, as expected from previous observa-tions by other authors and also the photometric studydescribed above. The large number of measured red-shifts allows an estimate of the velocity distribution ofthe galaxies in the field of NGC1600 to be determined,together with an estimate of the mass. Figure 3 showsthe relative velocity distribution of the 30 companionswith respect to NGC1600. Also shown is the best-fitGaussian, with a mean relative velocity of −
85 km s − and a velocity dispersion of 435 km s − .In order to obtain a more precise and less model-dependent measurement of these parameters we have alsoused the ROSTAT code to estimate the mean redshiftof the group and its radial velocity dispersion (Beers,Flynn & Gebhardt 1990), with the usual cosmologicalcorrection and the correction for velocity errors given byDanese, de Zotti & Tullio (1980). Given that there is alarge number of redshifts available, the biweight estima-tors were used for both the location and scale (Beers etal. 1990). Errors were obtained in all cases by jackniv-ing of the biweight. Using this machinery we measurethe group’s central location at 4634 ±
79 km s − , with aradial velocity dispersion value of 429 ±
57 km s − .GC1600 5 TABLE 1Galaxies in the NGC1600 area observed in this work (upper part) or in the literature.
RA (J2000) Dec (J2000) B mag v Error Comments04 : 31 : 39 . −
05 : 05 : 09 . ± a , 4739 ± b
04 : 30 : 04 . −
05 : 00 : 35 . . −
05 : 21 : 16 . . −
04 : 53 : 33 . . −
04 : 52 : 13 . . −
05 : 13 : 56 . . −
05 : 01 : 09 . . −
04 : 57 : 03 . . −
05 : 16 : 28 . . −
05 : 03 : 37 . ± c
04 : 31 : 43 . −
04 : 59 : 14 . . −
05 : 16 : 07 . . −
05 : 05 : 39 . ± d
04 : 31 : 58 . −
05 : 22 : 11 . ± e
04 : 32 : 03 . −
05 : 01 : 57 . . −
05 : 08 : 12 . . −
05 : 15 : 50 . . −
05 : 11 : 18 . . −
05 : 09 : 18 . . −
05 : 10 : 44 14.35 4267 25 NGC158004 : 28 : 25 . −
04 : 37 : 50 >
15 4680 6004 : 28 : 26 . −
04 : 33 : 49 15.75 4846 3904 : 30 : 34 . −
05 : 31 : 54 17.72 4908 6004 : 30 : 51 . −
05 : 47 : 54 13.72 4329 4 NGC159404 : 31 : 12 . −
05 : 31 : 44 16.26 5202 6004 : 31 : 38 . −
04 : 35 : 18 14.36 4008 8 NGC159904 : 31 : 52 . −
05 : 45 : 25 15.47 4121 4104 : 32 : 03 . −
04 : 27 : 38 14.51 4255 3 NGC160704 : 32 : 08 . −
04 : 12 : 43 15.74 4564 6004 : 33 : 01 . −
04 : 11 : 19 16.82 4904 1504 : 33 : 05 . −
04 : 17 : 51 14.40 4261 15 NGC1611
Note . — At the lower part of the table, we list other galaxies within a projected separation of 1 Mpc and 1000 km s − of NGC1600but not observed by us. Data from HyperLeda. a NED, from Collobert et al. 2006. b HyperLeda average. c NED, from de Vaucouleurs et al. 1991. d NED, from Simien and Prugniel 2000. e NED, from Huchra et al. 1993.
The redshift of NGC1600 lies 1 σ away of the systemcentral velocity, and thus it appears that NGC1600 isnot at the dynamical centre of the group, even though itis the brightest member by over 2 magnitudes. This isin agreement with Sivakoff et al. (2004) who found thatthe X-ray emission is centred slightly to the north-east ofNGC1600, also suggesting that the galaxy is not at thecentre of the gravitational potential. An estimate of thegeometry of the system can be ascertained from Figure4, where we show the positions and velocities relative toNGC1600 of the galaxies for which a redshift has beenmeasured.On the other hand, the velocity dispersion of 429km s − would imply a bolometric X-ray luminosity L X =2 . × erg s − , as in the empirical relation of Ortiz-Gilet al. (2004), which was measured from a sample of 171clusters drawn from the REFLEX catalogue. Sivakoff etal. (2004) measure an X-ray flux which is two orders ofmagnitude smaller . Although the L X − σ relation is notwell defined, NGC1600 is more than 3 σ away from the REFLEX luminosities are bolometric, while Sivakoff et al.(2004) measure only in the soft band. However, the correctionsfrom soft to bolometric represent a factor between 1.1 and 2 forthis kind of groups/clusters (Bohringer et al, 2004, Table 5). expected value, and hence lies significantly outside thescatter.A velocity dispersion value of 429 km s − for an el-liptical galaxy is very high, placing it well off the Faber-Jackson relationship (Faber and Jackson 1976). Hau andForbes (2006) measured the radial distribution of the ve-locity dispersion for the Reda et al. (2004) sample ofisolated galaxies. They find no galaxy with a similar ve-locity dispersion to that found here, with an increase inthe V /σ with radius and a general decrease of velocitydispersion with radius, out to the effective radius. Ve-locity dispersions of loose groups are found to be of theorder of a few hundreds of kilometers per second (e.g.Ramella et al. 1995), with a value of 429 km s − be-ing at the upper limit for loose groups, but at the lowerrange for clusters of galaxies. Making the somewhat ar-bitrary assumption that the group can be approximatedby an isothermal sphere it is possible to derive an upperlimit for the mass of the group from M ( R ) = ( πσ R ) /G (Binney and Merrifield, 1998). A simple application ofthis formula to the NGC1600 ‘group’ gives a mass of ap-proximately 2 × M ⊙ , typical of the richer groups orpoorer clusters. Although only 30 satellites have beendetected, we can also use the estimate of Bahcall and Smith, R.M. et al. -1000 -500 0 500 1000012345 Relative Velocity (km/s) Fig. 3.—
Histogram showing the velocity distribution of thegalaxies in the NGC1600 group, relative to NGC1600 itself. Thedashed line corresponds to the best-fit Gaussian, with v cent = −
85 km s − and σ v = 435 km s − . Tremaine (1981) to derive an estimate of the mass ofthe central object. Assuming that the galaxies are dis-tributed uniformly on radial and circular orbits, the massof the central object is given by M = 48 h r ∆ v / G i /π .Applying this formula to the group data gives a massestimate of 1 . × M ⊙ , in agreement with the isother-mal sphere determination. Zaritsky and White (1994)have shown from extensive modelling that errors in suchan estimate are not large as long as interlopers are ex-cluded. However, the presence of interlopers can havea serious effect on mass determinations of galaxies andclusters (e.g. Chen et al. 2006). To estimate the effect ofinterlopers on our sample, we can remove from the sam-ple those galaxies whose relative velocities are greaterthan 900 (500)kms − . This would drop the value of themass from 1 . × M ⊙ to 1 . × (7 . × ) M ⊙ .Further elimination of those galaxies at projected dis-tances greater than 1Mpc from NGC 1600 would dropthe mass estimate even further to 2 . × M ⊙ , butfor a sample of only 13 galaxies. This lower estimate istypical of loose groups (e.g. Parker et al. 2005) and isperhaps larger than that expected from dynamical ob-servations of the inner regions of elliptical galaxies suchas found from the study of planetary nebulae (e.g. Ro-manowsky et al. 2003). Fuzakawa et al. (2006) have alsoused the X-ray emission to derive an estimate of 10.58for the mass-to-light ratio for the NGC 1600 ‘group’ outto the effective radius of 13.8 kpc, giving a mass of ap-proximately 10 M ⊙ at this distance. Extrapolating thisestimate to larger radii is subject to considerable errorbut, if the radial mass-to-light ratio profile for NGC 1600is similar to that found for other ellipticals by Fuzakawaet al. (2006), the mass determination from the X-rayemission is not in major disagreement with that fromour dynamical study. However, the variations associated with the different methods suggest that our estimatesof the group mass are still subject to an uncertainty ofabout a factor of two.The results of the number density and radial veloc-ity studies allow us to probe further our definition ofan isolated galaxy. The former study indicates thatthere is a population of much fainter galaxies surround-ing NGC1600, with a small number of brighter galaxiesin the immediate vicinity which are not seen in the sam-ple of Reda et al. (2004). However, these bright galax-ies have a negligible effect on the overall number densityprofile around NGC1600. The profile is therefore not sig-nificantly different from that found from other isolatedearly-type galaxies. The published photometric proper-ties are also similar to those for isolated ellipticals, withlittle evidence of recent merging. The dynamical studyhowever, together with the X-ray data, indicates thatNGC1600 is not sitting at the centre of the potential andis surrounded by a massive halo extending out to severalhundred kiloparsecs. This would normally be taken asevidence that NGC1600 is a member of a group of galax-ies.In any case, a simple comparison of the ( J − K ) vs K colour diagram of the NGC1600 group of galaxies withthat of the Coma Cluster (Figure 5) shows that thereare similarities between both fields. The main differ-ences are, of course, the large difference in richness andthe already mentioned magnitude gap between NGC1600and the second K -band brightest galaxy in the group(NGC1611), which is much larger than in Coma.There are several similarities between NGC1600 andthe fossil groups of Jones et al. (2003): the large magni-tude difference between it and its neighbours, the pres-ence of a surrounding faint population, and the presenceof extended X-ray emission. In addition, the absolutemagnitude of NGC1600 ( M R ∼ − .
0) lies within therange of luminosities found for other fossil group centralgalaxies, although at the fainter end of the range. More-over, the measured velocity dispersion and absolute B -band magnitude position NGC1600 perfectly in the σ V vs M B relation for fossil groups, as described by Khos-roshahi et al (2006). If the scaling properties of the fossilgroups as measured by D’Onghia et al (2005) in theirsimulations apply, NGC1600 would have been already inplace (that is, more than 50% of the mass would alreadyhave been assembled) 7-8 Gyrs ago, which matches theavailable age estimates. However, the X-ray propertiesof the NGC1600 group are very different from those ex-pected, with a much smaller extent and a much lowerluminosity (Sivakoff et al 2004). Thus, NGC1600 couldbe taken as a new class of fossil group, where the centralgalaxy has grown from the merger of several fainter groupmembers but is lacking the relatively intense X-ray emis-sion normally associated with loose groups. Thus it couldbe the result of a merger of a poor group, where the X-rayemission is expected to be much weaker. The presenceof dust and emission regions within the galaxy supportthis merger hypothesis. NGC 1132, an isolated ellipticalgalaxy studied by Mulchaey and Zabludoff (1999), hassimilar properties to NGC 1600, with a faint extended X-ray envelope and surrounding excess population of dwarfgalaxies, although the dynamical properties of the groupare uncertain. Thus NGC1600 may not be the only ex-ample of such a past merger.GC1600 7 Fig. 4.—
UKST Sky Survey 1 degree square image of the NGC1600 field. Those galaxies for which a redshift has been measured aremarked, and their velocity relative to NGC1600 has been labeled. Notice that some galaxies in Table 1 are outside this field. Use of theseimages is courtesy of the UK Schmidt Telescope (copyright in which is owned by the Particle Physics and Astronomy Research Councilof the UK and the Anglo-Australian Telescope Board) and the Digitized Sky Survey created by the Space Telescope Science Institute,operated by AURA, Inc, for NASA, and is reproduced here with permission from the Royal Observatory Edinburgh.
Fig. 5.— JK colour-diagram of the NGC1600 field. The darklarge circles correspond to objects with redshifts that link themto NGC1600, whereas the small circles are field objects withouta redshift. As a comparison, the crosses mark the positions ofgalaxies in the Coma cluster field, corrected to the distance ofNGC1600. CONCLUSION
Although NGC1600 is clearly surrounded by a num-ber of other galaxies, the large magnitude difference sug-gests that it can be treated as an isolated, field, ellip-tical. Analysis of the APM scans of UKST Sky Surveyplates shows that there is a clear excess of galaxies withinthe neighbourhood of NGC1600. In partial agreementwith Reda et al. (2004) the excess is most clearly evi-dent for the fainter galaxies, with galaxies brighter than M R = −
16 not showing such a clear excess. The num-ber of galaxies, together with their number density pro-file, around NGC1600 is in agreement with the results ofSMG indicating that this galaxy is not unique amongstisolated early-type galaxies.Velocity measurements of galaxies within the field showa large excess at similar redshifts to NGC1600. The dy-namics of these galaxies indicates that the mass of thegroup is at least 10 M ⊙ , a value typical of poor groups.Combining the data from these and previous studies in-dicates that NGC1600 has many similarities to the fossilgroups first detected in X-rays, although there are sev-eral differences, most notably in the X-ray properties.We therefore suggest that NGC1600 is a new type of fos-sil group, the result of the merger of a loose group ofgalaxies, with the last merger occuring over 4 Gyr ago.The similarity between NGC1600 and other galaxiesin the SMG sample of isolated ellipticals suggests that Smith, R.M. et al.many such objects may be the result of the merger ofloose groups. This will only be determined by a detailedphotometric and dynamical study of a larger sample ofisolated galaxies. As the majority of galaxies lie in loosegroups, including our own, such an investigation wouldlead to a greater understanding of galaxy evolution.The authors acknowledge the comments of our anoy-mous referee, which helped in clarifying the content ofthis paper. We also thank Elena D’Onghia for her helpfulcomments. VJM, AFS, FJB, and AOG acknowledge sup-port from the Spanish Minsiterio de Educaci´on y Cien-cia (MEC) through project AYA2006-14056 (includingFEDER). AFS acknowledges support from the SpanishMEC through a “Ram´on y Cajal” contract. We thankVicent Peris for his help in processing the DSS image of NGC1600 for Fig. 4 using PixInsight. This work is basedon observations obtained at the 4.2-m William HerschelTelescope (WHT), operated on the island of La Palma bythe Isaac Newton Group in the Observatorio del Roquede Los Muchachos of the Instituto de Astrof´ısica de Ca-narias. The authors acknowledge the assitance givenby the support astronomers at WHT. This publicationmakes use of data products from the Two Micron AllSky Survey, which is a joint project of the University ofMassachusetts and the Infrared Processing and AnalysisCenter/California Institute of Technology, funded by theNational Aeronautics and Space Administration and theNational Science Foundation. We acknowledge the usageof the HyperLeda database (http://leda.univ-lyon1.fr). Facilities: