Discovery of an extremely gas-rich dwarf triplet near the center of the Lynx-Cancer void
aa r X i v : . [ a s t r o - ph . C O ] O c t Mon. Not. R. Astron. Soc. , 1–9 (2012) Printed 13 April 2018 (MN L A TEX style file v2.2)
Discovery of an extremely gas-rich dwarf triplet near thecenter of the Lynx-Cancer void
J.N. Chengalur, ⋆ S.A. Pustilnik , National Centre for Radio Astrophysics, Post Bag 3, Ganeshkhind, Pune 411 007, India Special Astrophysical Observatory of RAS, Nizhnij Arkhyz, Karachai-Circassia 369167, Russia Isaac Newton Institute of Chile, SAO branch, Nizhnij Arkhyz, Russia
Accepted 2012 September ??. Received 2012 August 23
ABSTRACT
Giant Metrewave Radio Telescope (GMRT) H i observations, done as part of anongoing study of dwarf galaxies in the Lynx-Cancer void, resulted in the discovery ofa triplet of extremely gas rich galaxies located near the centre of the void.The tripletmembers SDSS J0723+3621, J0723+3622 and J0723+3624 have absolute magnitudes M B of –14.2, –11.9 and –9.7 and M (H i )/ L B of ∼ ∼
10 and ∼
25, respectively. Thegas mass fractions, as derived from the SDSS photometry and the GMRT data are 0.93,0.997, 0.997 respectively. The faintest member of this triplet SDSS J0723+3624 is oneof the most gas rich galaxies known. We find that all three galaxies deviate significantlyfrom the Tully-Fisher relation, but follow the baryonic Tully-Fisher relation. All threegalaxies also have a baryon fraction that is significantly smaller than the cosmic baryonfraction. For the largest galaxy in the triplet, this is in contradiction to numericalsimulations. The discovery of this very unique dwarf triplet lends support to the ideathat the void environment is conducive to the formation of galaxies with unusualproperties. These observations also provide further motivation to do deep searches ofvoids for a “hidden” very gas-rich galaxy population with M B & − Key words: galaxies: dwarf – galaxies: evolution – galaxies: individual: SDSSJ0723+3621, J0723+3622, J0723+3624 – galaxies: kinematics and dynamics – radiolines: galaxies – cosmology: large-scale structure of Universe
Early redshift surveys established that the spatial distribu-tion of bright galaxies is highly inhomogeneous and that theproperties of the galaxy population varies with environment.The spatial distribution was found to consist of large underdense regions (“voids”) surrounded by galaxies in sheets andwalls (J¨oeveer, Einasto & Tago 1978; Kirshner et al. 1981;Geller & Huchra 1989). Further, the morphological mix ofgalaxies was found to vary systematically with galaxy den-sity. The fraction of late type galaxies monotonically in-creases as one goes from high density to low density re-gions (Postman & Geller 1984). Subsequent numerical simu-lations showed that the existence of voids can be understoodas a consequence of biasing in the formation of galaxy halos,with the most massive halos being formed in regions of highdensities (White et al. 1987). However, numerical simula-tions also predicted that the voids should be filled with small ⋆ E-mail: [email protected] (JNC), [email protected] (SAP) mass halos (e.g. Davis et al. 1985; Gottl¨ober et al. 2003).Peebles (2001) pointed out that, contrary to this expec-tation, the known dwarf galaxies instead follow the samelarge scale distribution as the bright galaxies, a discrepancythat he dubbed the “void phenomenon”. We note that ear-lier (Pustilnik et al. 1995) as well as recent (Kreckel et al.2012) studies do find dwarf galaxies inside voids, howeverit remains true that the brighter dwarfs generally lie nearthe void walls. Peebles (2001) also highlighted that if thesmall dark halos produced in voids preferentially fail to hostgalaxies, this would correspond to a discontinuous change ingalaxy properties with density. This distinguishes the “voidphenomenon” from the observed “morphology-density” re-lationship in which the morphological mix varies smoothlywith density. Observational and theoretical studies of thevoid galaxy population have since been largely focused onthese two (possibly related) issues viz. (i) a search for the“missing” dwarfs in voids and (ii) the influence of the largescale environment on galaxy properties.Regarding the issue of missing dwarfs, deep searches c (cid:13) J.N. Chengalur, S.A. Pustilnik of voids have shown that they do not contain dwarfgalaxies in the numbers predicted by simulations (seee.g. Tikhonov & Klypin 2009; Kreckel et al. 2012). Thereason for this discrepancy is unclear, although therehave been numerous suggestions that the formation ofgalaxies in small dark matter halos in voids is sup-pressed (e.g. Tinker & Conroy (2009); see also Kreckel et al.(2011)). Essentially, if small halos in voids are baryon de-ficient this would resolve the problem of missing dwarfs.Hoeft and Gottl¨ober (2010) examined the baryon fractionof small halos in both voids and filaments using high reso-lution simulations, and found no dependence of the baryondeficiency on environment. The discrepancy between numer-ical simulations and the observations hence remains a puzzle.Regarding the issue of the effect of the large scale en-vironment on the properties of void galaxies, studies usingSDSS selected samples established that the void galaxy pop-ulation is significantly bluer and has a higher star forma-tion rate as compared to the high density galaxy popula-tion (Rojas et al. 2004, 2005). However Patiri et al. (2006)(see also (Park et al. 2007)) show that this difference is al-most entirely due to the morphology-density relation. Latetype galaxies which are more dominant in low density re-gions, have bluer colours and higher star formation ratesthan early type galaxies. At a fixed luminosity and mor-phology the properties of the detected void galaxies are sta-tistically identical to that of galaxies in dense regions. Itis worth noting here that these conclusions relate only tothe upper part of the whole luminosity (or mass) range ofvoid galaxies ( M B , r . −
16 mag) and do not include studyof possible differences in parameters such as the gas phasemetallicity and gas mass fraction.The gas mass fraction of void galaxies was studied inearlier works which looked at the distribution of M (H i )/ L B .Huchtmeier, Hopp, & Kuhn (1997) found that dwarf galax-ies closer to the center of the void had a higher M (H i )/ L B than galaxies close to the walls. Similarly, Pustilnik et al.(2002) found marginal evidence for low luminosity galaxiesin voids to have a higher M (H i )/ L B than those in higherdensity regions. Extrapolation of the obtained trends to therange M B > −
15 indicated that for lower mass dwarfs thedifference could be higher. On the other hand, Kreckel et al.(2012) show that the H i gas content of void galaxies arestatistically indistinguishable from galaxies in filaments andwalls, at least for galaxies brighter than M r ∼ −
16 mag.This result does not contradict earlier results, and under-lines the need of deeper void galaxy samples.A possible resolution of the discrepancy between thepredictions of the numerical simulations and the observa-tions is that the dwarfs predicted to exist in voids are fainterthan what the observations have probed so far. For exam-ple, in their simulation, Kreckel et al. (2011) find that whileluminous dwarfs ( M r brighter than ∼ −
18 mag) in voids arestatistically indistinguishable from similar dwarfs in higherdensity regions, fainter dwarfs ( M r ∼ −
16 mag) are signif-icantly bluer and have higher specific star formation ratesthan their higher density counter parts. They also find asignificant excess of faint dwarf ( M r ∼ −
14 mag) galaxiesthat are preferentially located in low density regions nearthe void centre. To complement their numerical simulationsKreckel et al. (2012) used the SDSS to identify a popula-tion of void galaxies with M r > − . Table 1.
Parameters of the GMRT observationsJ0723+36 tripletDate of observations 2011 Nov 25Field center R.A.(2000) 07 h m s Field center Dec.(2000) +36 o ′ ” Central Velocity (km s − ) 950.0Time on-source (h) ∼ − ) ∼ ) 40 × − ) 2.8 study still fainter objects one needs to focus on the nearbyvoids.In a recent series of papers (Pustilnik & Tepliakova(2011) (Paper I), Pustilnik, Tepliakova & Kniazev (2011)(Paper II), Pustilnik et al. (2011c) (Paper III)) a sampleof 79 galaxies residing in the nearby Lynx-Cancer void waspresented. The sample galaxies have M B down to -12 mag,with reasonable completeness level at M B ∼ − . ∼ − ) is about one order of magnitude smallerthan the mean value for the faint SDSS galaxies derived byBlanton et al. (2005). More than half of the Lynx-Cancervoid sample consists of low surface brightness dwarfs (LS-BDs). Measurements of O/H are available for ∼
60% of thesample, and shows that the metallicity of the void galaxiesis on average ∼
30% lower than that of their counterpartsin denser regions. About 10% of the sample are deficient inmetals by factors of 2-7. A GMRT based H i study of thesedwarfs is in progress. Here we report on a highly unusualsystem found in the course of the H i observations, viz. anextremely gas rich triplet of LSBD galaxies, located insidethe central 10% of the void volume. GMRT H i ′′ resolution.In Fig. 1[A] is shown the GMRT H i map of theJ0723+36 system. At the time of the observations, onlytwo galaxies, viz. SDSS J0723+3621, and SDSS J0723+3622were known to lie within the observed data cube. At theadopted distance of 16 Mpc (see below) to this group the sep-aration of the pair is 12.1 kpc. As can be seen from the figure,the GMRT observations detected one more H i source, whichcorresponds to the galaxy SDSS J0723+3624. The projectedseparation between this companion and the brighter of thetwo galaxies in the pair (viz. J0723+3621) is 23.9 kpc. SDSS g band images showing these galaxies is shown in Fig. 2and Fig. 3 (a colour composite for this pair is also shown c (cid:13) , 1–9 xtremely gas rich dwarf triplet in the Lynx-Cancer void Figure 1. [A] The integrated H i emission (moment0 map) fromthe J0723+36 system. The angular resolution is 40 ′′ . The con-tours start at 3 × atoms cm − and are in steps of 1.414.[B] The velocity field (moment1 map) of the J0723+36 system,derived from the 40 ′′ resolution data. The iso-velocity contoursstart at 870.0 km s − and are in steps of 6 km s − . Figure 2.
The SDSS g band image showing the main two galax-ies in the triplet, viz. J0723+3621 (the edge on galaxy) andJ0723+3622 (the fainter companion). Figure 3.
The SDSS g band image showing the faintest mem-ber of the triplet, J0723+3624. The GMRT H i contours are alsooverlayed, the contour levels are the same as in Fig. 1. in Paper III). The two brighter galaxies, J0723+3621, andJ0723+3622 are clearly interacting, with a bridge of H i con-necting them. For the third much smaller system there is ahint of an extension to the North-West, but the resolution ismarginal. The velocity field (with isovelocity lines in steps of6.0 km s − ) for the whole system is shown in Fig. 1[B]. Allthree galaxies show clear velocity gradients, although in allcases the velocity field is disturbed. In the case of the galaxypair, this is clearly due to the ongoing tidal interaction. Thespins of both components of the pair are aligned with theirorbital angular momentum, consistent with the pair under-going a prograde collision. A prograde encounter geometryis also consistent with the significant tidal distortions seen.A continuum image made using all the available chan-nels shows no emission from the triplet galaxies. The rmsnoise level is ∼ . ′′ .The main parameters of the three galaxies in the0723+36 system are given in Tab. 2. The H i parametersare derived from the integrated single dish profiles shown inFig 4. Because the velocity fields are disturbed, we do notattempt to derive rotation curves. Instead we use the veloc-ity widths obtained from the integrated profiles to estimatethe dynamical mass related quantities. The optical parame-ters are derived from the SDSS DR7 data (Abazajian et al.2009). In the case of the newly discovered companion galaxyJ0723+3624, there are two extended SDSS objects, sepa-rated by ∼ ′′ (0.4 kpc in projection) seen near the centerof the H i emission. We refer below to these two objects asso-ciated with the companion J0723+3624 as the NE and SWcomponents. The NE component (J072320.57+362440.8),has g =21.42 and somewhat blue colours, (( g − r ) = 0 . ± . g =21.29 ± g − r ) = 0 . ± .
08. TheSW component J072320.32+362436.7 is ∼ g -filter and is significantly redder (( g − r ) = 0 . ± . g -flux that is ∼ g − r ) = 0 . ± .
11. This ‘red nebulosity’ looks similarto the very faint red galaxies seen to the North and East of c (cid:13) , 1–9 J.N. Chengalur, S.A. Pustilnik
800 900 1000 11000204060 800 900 1000 1100-10010203040 800 900 1000 1100-100102030
Figure 4.
Integrated H i spectra for the members of the J0723+36 triplet. The spectra have been derived from the 40 ′′ resolution datacube. The left panel is for J0723+3621. Note that the first profile has a double horned shape, characteristic differentially rotating disks.According to Geha et al. (2006), 18% of dwarfs with M r > −
16 have such profiles. The middle panel is for J0723+3622 while the rightpanel is for J0723+3624. the blue component, which could represent a small group ofdistant galaxies. We hence assume that the red nebulosity isunrelated to the H i -cloud. In any case, including this com-ponent would make only a minor difference to the total flux– making the object brighter in the B -band by 0.26 mag.The rows of Tab. 2 are as follows: the J2000 RA andDec; A B , the Galactic extinction in B -band; the total B -magnitude, not corrected for A B , obtained by transforma-tion from the total g and r , according to the formulae givenin Lupton et al. (2006); the ( g − r ) colour; the ( B − V )colour, computed from the observed ( g − r ) and for the PE-GASE2 constant evolutionary tracks with z = 0 .
002 (or Z = Z ⊙ /10). the heliocentric velocity, obtained from H i -profile in this paper; the adopted distance, accounting forthe updated V hel and the large negative velocity correction,described in Paper I; the calculated absolute blue magni-tude, optical sizes (angular and linear), corrected for theGalactic extinction and inclination; the central inclinationcorrected surface brightness in B -band; the measured H i -flux in units of Jy km s − ; the profile widths W and W ; the inclination angle estimated from the SDSS im-ages; the H i mass M (H i ); the total baryonic mass com-puted as M bary = 1 . × M (H i ) + M star . M star is the stellarmass computed from the SDSS data using the total mag-nitude in g -filter and the total ( g − i ) colour, corrected forthe Galactic extinction (from NED, following to Schlegel etal. 1998), and the mass to luminosity ratio Υ determinedfrom the prescription given in Zibetti et al. (2009) (simi-lar to used in Paper III); the total dynamical mass, com-puted using the formula M dyn = 2 . × × R kpc × V − ,where R kpc is the radius measured from the H i images atthe 3 × atoms cm − level, and V km s − is the rota-tional velocity computed from W after correction for incli-nation and turbulent motions using the prescription givenin Verheijen & Sancisi (2001); M vir , the virial mass, com-puted from the circular velocity estimated as above, and theformulae given in Hoeft et al. (2006); R vir , the virial radius,computed from the circular velocity estimated as above, andthe formulae given in Hoeft et al. (2006); the ratio of H i mass to blue luminosity, M (H i )/ L B in solar units; the gas Table 2.
Main parameters of the J0723+36 tripletParameter J0723+3621 J0723+3622 J0723+3624R.A.(J2000.0) 07 23 01.42 07 23 13.46 07 23 20.57DEC.(J2000.0) +36 21 17.1 +36 22 13.0 +36 24 40.8 A B (from NED) 0.23 0.23 0.23 B tot ± ± g − r ) , tot ± ± ± B − V ) , tot ± ± ± V hel (H i )(km s − ) 917 ± ± ± M –14.24 –11.94 –9.68Opt. size ( ′′ ) × (2) × × × (2) × × µ (mag arcsec − ) 24.14 24.36 24.6:H i int.flux 3.74 ± ± ± W (km s − ) 100.2 ± ± ± W (km s − ) 122.5 ± ± ± i (degrees) 90: 60: 60: V rot (H i )(km s − ) 54.0 35.2 15.3 M (H i ) (10 M ⊙ ) 22.6 9.61 2.9 M bary (10 M ⊙ ) 32.37 12.8 3.86 M dyn (10 M ⊙ ) 623.8 132.5 16.91 M vir (10 M ⊙ ) 359.5 102.3 8.9 R vir (kpc) 87.3 57.4 25.4 M (H i )/ L B f gas f bar mass fraction f gas = 1 . × M (H i ) /M bary ; the baryon fraction f bar = M bar /M vir . As can be seen from Tab. 2, all three galaxies have very large M (H i )/ L B ratios; in fact J0723+3624 has one of the largesratios known. The corresponding gas mass fractions arealso extremely large. Even if we assume that the colours ofJ0723+3622 and J0723+3624 are redder by 1 σ that the mea-sured values, their gas mass fractions remain & c (cid:13) , 1–9 xtremely gas rich dwarf triplet in the Lynx-Cancer void Figure 5.
Comparison of the M (H i )/ L B ratio of the galaxiesin the Lynx-Cancer triplet (filled squares), with galaxies fromthe FIGGS sample (Begum et al. 2008b, crosses). Three very gasrich dwarfs from the Void Galaxy Sample (VGS) of Kreckel et al.(2012) are shown as circles, and data for 6 other extremely gasrich galaxies (see Tab. 3) are shown in triangles. The solid line isthe best fit relation for the FIGGS galaxies. forming galaxies the gas mass fraction is known to increasewith decreasing luminosity (see e.g. McGaugh & de Blok1997; Geha et al. 2006). We show in Fig. 5 M (H i )/ L B for the FIGGS ((Begum et al. 2008b)) sample, which alsoclearly shows this trend. The data for the galaxies from theJ0723+36 triplet are also shown, and, as can be seen, fortheir given luminosities, all three galaxies lie at the extremegas rich end of the distribution.We also compare the galaxies location in the Tully-Fisher (TF) and Baryonic Tully-Fisher (BTF) diagrams. Wenote that the inclinations estimated for these galaxies aresomewhat uncertain, however this should affect the TF andBTF relations equally. In Fig. 6[A] is shown the Tully Fisherrelation for the FIGGS galaxies. Also overplotted is the TFrelation for bright galaxies, as determined by Tully & Pierce(2000). As expected for dwarf galaxies, the FIGGS galax-ies are underluminous for their velocity width ( see alsoBegum et al. 2008a). Once again, the triplet galaxies fallat the extreme end of the distribution. Fig. 6[B] shows thebaryonic Tully-Fisher relation, with for reference the BTFrelation shown from De Rijcke et al. (2007). Despite beingextremely gas rich, the triplet galaxies do follow the BTFrelation.The baryon fraction of small galaxies is also an inter-esting quantity to compare against model predictions. Forour extremely gas rich galaxies, the baryon fraction can beaccurately measured. From numerical models (Hoeft et al.2006; Hoeft and Gottl¨ober 2010), one would expect thatgalaxies with circular velocities &
50 km s − would havea baryon fraction equal to that of the cosmic value of ∼ . − , the predicted baryon fraction is an orderof magnitude below the cosmic value. As can be seen fromTab. 2, for all galaxies the baryon fraction f bar is more thanan order of magnitude lower than the cosmic baryon frac-tion. In the case of the brightest galaxy J0723+3621, thebaryon fraction is ∼ /
20 that of the cosmic baryon frac-tion, while one would expect it to have the cosmic baryonfraction. This galaxy is clearly edge on, and this result ishence unlikely to be due to an uncertain inclination an-gle. McGaugh et al. (2010) has earlier highlighted that themeasured baryon fraction for galaxies with rotation veloc-ities in this range is significantly smaller than the cosmicbaryon fraction. In this respect, J0723+3621 is similar todwarf galaxies located outside of voids. Interestingly, f bar appears to decrease with increasing velocity width, i.e. thereverse of what is predicted. One possible reason for thiscould be that for the smaller galaxies the baryons do notsample the flat part of the rotation curve, and hence W underestimates the circular velocity. It is also worth notingthat the expected virial radius of even the smallest galaxiesdark matter halo is larger than the separation of the galax-ies in this triplet. If one regards the entire triplet as a singlesystem then the f bar is ∼ .
01, about 16 times smaller thanthe cosmic baryon fraction.The triplet of galaxies that we are discussing here arefound in the inner 10% volume of the Lynx-Cancer void.Kreckel et al. (2012) also find 3 similarly gas rich galaxies(VGS 7a, VGS 9a, VGS 12) in their void survey. Their abso-lute blue magnitudes computed from the SDSS magnitudesfollowing the same procedures as for our triplet galaxies are M B of –13.86, –11.17 and -16.49 respectively. The corre-sponding M (H i )/ L B ratios are is ∼ ∼ ∼ M (H i )/ L B & D ∼ ∼ c (cid:13) , 1–9 J.N. Chengalur, S.A. Pustilnik
Figure 6. [A] : The Tully-Fisher relation: W vs M B . The reference TF relation line (solid) for brighter galaxies (Tully & Pierce 2000). [B]: The Baryonic Tully-Fisher relation. The solid line shows the BTF relation from De Rijcke et al. (2007). The symbols are the sameas those used in Fig. 5 projected separation of only 150 kpc from the large spiralgalaxy NGC 1376 (Pustilnik et al. 2001). Peebles (2001) ex-amined its location with respect to galaxies from the ORS(Santiago et al. 1995), and concluded that while it is not lo-cated in a dense region, neither is it located in a void. Insummary, while all of the gas rich galaxies appear relativelyisolated, it is not the case that they are all located deepinside voids.Next, we use star formation models to constrain the evo-lutionary history of these extremely gas rich galaxies. For 5of these galaxies we only have the ( B − V ) colour available,and we hence use this colour to constrain the star forma-tion history. Unfortunately, due to the degeneracy of theevolutionary tracks for continuous and instantaneous starformation laws in B − V, V − R diagrams one can not distin-guish between ∼ ∼ u or U band data in the analysisallows one in many cases to do this. For example Pustilniket al. (2008, 2010, 2011b) use SDSS u, g, r colours for sim-ilar gas-rich metal-poor blue galaxies to show that in mostcases a continuous star formation law is preferable. Assum-ing that the LSB galaxies in the current sample have beenforming stars at a constant rate, we used PEGASE2 mod-els to compute the dependence M (H i )/ L B and the B − V colour on the age of the galaxy. We show in Fig. 7 a grid ofmodels with f gas =0.5, 0.8, 0.9, 0.95, 0.98 and 0.99, and starformation which started 0.5, 1, 3, 5, 12 Gyr ago. Consistentwith the mass to light ratios that we have adopted in Tab. 2one can see that models with f gas . .
95 do not match theobserved M (H i )/ L B and B − V colours. Further, for the“blue” colours ( B − V < +0.25), typical of the gas-rich LS-BDs considered here, the corresponding ages are < f gas & i -survey ALFALFA (Haynes et al. 2011) has a signifi-cantly higher sensitivity and angular resolution than pre-vious surveys. However, even in the ALFALFA survey, ob-jects like the faintest member of our triplet would be diffi-cult to detect outside the Local Volume. J0723+3624 with F (H i )=0.48 Jy km s − would fall below the survey detec-tion limit were it to be placed ∼ D centre .
20 Mpc. De-tection of substructure in systems like this triplet wouldhowever require follow up synthesis imaging observations.We note in this context that Kreckel et al. (2012) find thatvoid dwarfs show similar small scale clustering as dwarfs indenser environments. For distances
D >
16 Mpc (distancemoduli µ >
31 mag) these faint LSBDs with B tot >
19 willnot be easily identified, either via blind H i -surveys, nor viarecently conducted wide-field spectral surveys, like the SDSSor 2dFGRS. The existence in voids of ’unknown’ populationof very gas-rich LSB dwarfs with M B & –11 which escapeddetection in previous studies, hence remains a viable op-tion. Systematic searches for such faint dwarfs will requirethe next generation optical and radio surveys. c (cid:13) , 1–9 xtremely gas rich dwarf triplet in the Lynx-Cancer void Table 3.
Main parameters of galaxies with the highest M (H i )/ L B ratiosParameter And IV SBS 0335–052W HI 1225+01 SW HI 1225+01 NE UGCA 292 DDO 154R.A.(J2000.0) 00 42 32.30 03 37 38.40 12 26 55.00 12 27 46.29 12 38 44.63 12 54 05.20DEC.(J2000.0) +40 34 18.7 −
05 02 36.4 +01 24 35.0 +01 35 57.1 +32 45 01.5 +27 08 59.0 B tot hel (H i )(km s − ) 234 4017 1226 1299 308 374Distance (Mpc) 6.1 53.6 20.0 20.0 3.61 4.0M –12.6 –14.70 > –11.5 –15.49 –11.76 –14.13( B − V ) µ (mag arcsec − ) 23.3 22.5 > ∼ i int.flux (6) M (H i ) (10 M ⊙ ) 15.8 58.3 86 220 5.4 31.0 M (H i )/L B >
155 10.1 6.93 4.52 f gas > D =20 Mpc, while ∼
10 and ∼
15 Mpc are possible alternatives; (5) – data for UGCA 292: van Zee 2000, Makarova et al. 1998; (6) –data for DDO 154: Carignan & Freeman 1988, Carignan & Beaulieu 1989; Walter et al. 2008 (THINGS), O/H fromMoustakas et al. 2010.
Figure 7.
Constant star formation rate PEGASE2 models(dashed) of log( M (H i )/ L B ) versus ( B − V ) for galaxies with con-stant gas mass-fraction f gas . Model lines are shown for f gas of0.50, 0.80, 0.90, 0.95, 0.98 and 0.99. The lines connect pointscorresponding to ages of 0.5, 1.0, 3.0, 5.0 and 12.0 Gyr sincethe start of the star formation. The PEGASE2 models also as-sume a Salpeter IMF and z =0.002 ( Z ⊙ /10). M (H i ) is takento be 0.75 M gas , and L B is estimated from M (stars), with M/L =Υ(
B, B − V ) according to Zibetti et al. (2009). Squaresshow positions of the two Lynx-Cancer triplet dwarfs. Circles de-note the three VGS galaxies and triangles the sample of gas-richgalaxies listed in Tab. 3. In summary we report the discovery of an extremely gas richtriplet of galaxies near the centre of the nearby Lynx-Cancervoid. The triplet consists of the LSB galaxies J0723+3621( M B = –14.2), J0723+3622 ( M B = –11.9) and J0723+3624( M B = –9.7) which lie within a projected separation of ∼
25 kpc, and a radial velocity interval of ∼
55 km s − . The M (H i )/ L B ratios are ∼ M (H i )/ L B and blue colours of the fainter two members of this tripletare consistent with star formation that started relatively re-cently ( . ACKNOWLEDGEMENTS
SAP acknowledges the support of this work through theRFBR grants No. 10-02-92650-IND and 11-02-00261 andthe Federal Target Innovative Program under the contractNo.14.740.11.0901. The authors thank Y.Lyamina for pro-viding the independent photometry for the faint galaxiesprior publication. This paper used observations made usingthe GMRT, which is operated by NCRA-TIFR. The au-thors acknowledge the spectral and photometric data andrelated information available in the SDSS database used forthis study. The Sloan Digital Sky Survey (SDSS) is a jointproject of the University of Chicago, Fermilab, the Insti-tute for Advanced Study, the Japan Participation Group,the Johns Hopkins University, the Max-Planck-Institute for c (cid:13) , 1–9 J.N. Chengalur, S.A. Pustilnik
Astronomy (MPIA), the Max-Planck-Institute for Astro-physics (MPA), New Mexico State University, PrincetonUniversity, the United States Naval Observatory, and theUniversity of Washington. This research has made use ofthe NASA/IPAC Extragalactic Database (NED).
REFERENCES c (cid:13) , 1–9 xtremely gas rich dwarf triplet in the Lynx-Cancer void Walter F., Brinks E., de Block W.J.G., Bigiel F., KennicuttR.C., Thornley M.D., Leroy A., 2008, AJ, 136, 2563White S. D. M., Davis M., Efstathiou G., Frenk C. S., 1987,Nature, 330, 451Zackrisson E., Bergvall N., ¨Ostlin G., 2005, A&A, 435, 29Zibetti S., Charlot S. & Rix H.-W., 2009, MNRAS, 400,1181This paper has been typeset from a TEX/ L A TEX file preparedby the author. c (cid:13)000