arXiv:astro-ph/9909198v1 13 Sep 1999 fEOi aaa,drn h cec eicto Program Verification Science the during Paranal, in ESO of to ⋆ requests offprint Send Dwarf Piersimoni, M. A. the on investigation new A 4 3 2 1 oebr1 2018 1, November n h airto fteRB( RGB the of adopt- calibration By the investigations. previous ing in is suggested values determination the distance This Mpc. esi h bevdcut a h 2 the -at counts observed the in cess nei h airto fteTG ehdadi the in and method TRGB differ- the to of calibration due ( in the mainly and in are distance ence determinations the in in both the available disagreement evaluations The similar literature. than metal-rich more dex l fbih tr scnitn ihteocrec fa of occurrence the with consistent episode is formation sam- this bright that of confirms comparison stars, ple observations The blue and component. bright theory stellar of between young Red sample a well-defined to a a belonging and features: (RGB), key Branch (VLT). two Telescope Giant Large present Very CMDs ESO the The the on with camera collected I galaxy, data FORS (SV) dwarf Verification Antlia Science the on based of (CMDs) diagrams magnitude ealct fAti tla ouaini fteodrof order the of mean is the population that stellar [ estimated Antlia we of metallicity, metallicity the of function iha g prxmtl qa o07Gyr. 0.7 to component equal stellar approximately a age to due an be with suggest might age discrepancy stellar this on that luminosity TRGB the the on of arguments dependence Plain predictions. theoretical with pared TG)poie ySlrs&Csii(98 eesti- ( we is modulus (1998) distance Cassisi Antlia & the that Salaris mated by provided (TRGB) one. population old stellar the young than this concentrated that centrally find more also is et we Aparicio 1997), 1997; al. al. et (Sarajedini investigations previous 25 e/H F 6(31.;1.61 90.;1.71 95.;19.63.1) 19.53.1; 19.37.1; 19.06.1; 16.06.1; (03.11.1; 06 are: codes thesaurus later) Your hand by inserted be (will no. manuscript A&A V sevtroAtooiod ret,VaGB ipl 1 4 11, Tiepolo G.B. Via Trieste, M¨unc di bei Astronomico Mo Garching Osservatorio 00040 D-85748 Observatory, 33, Frascati Southern via European 64 Roma, Maggini, di M. Astronomico via Osservatorio Collurania, di Astronomico Osservatorio . 89 ae nosrain olce ihVTU1telescope VLT-UT1 with collected observations on Based − epeetde ( deep present We Abstract h ieeta G uioiyfnto hw nex- an shows function luminosity RGB differential The yaotn h e airto fteTpo h RGB the of Tip the of calibration new the adopting By ± I oo ne s ealct relation. metallicity vs. index color ) ] 0 − ≈ . 0mg n hrfr distance a therefore and mag, 10 1 . .Ti ealct siaei tlat0.3 least at is estimate metallicity This 3. 1 .Bono, G. ..Piersimoni A.M. : ≈ ,V I, 0 . y g.I gemn with agreement In ago. Gyr 1 − I 2 n ( and ) V .Castellani, M. − σ ≈ I oo ne sa as index color ) ee-we com- when level- ,B I, 3 agrthan larger 13% D 1 = m − − . I 51 color- ) 2 M ± .Marconi, G. ) 0 0 . 07 = 0 eao Italy Teramo, 100 11Tise Italy Trieste, 0131 e,Germany hen, t ozoCtn,Italy Catone, Porzio nte rbesee nteinrotrgosmk hscasof class this make crowding regions evi- by innermost the the affected in and marginally even telescopes only area problems are collecting 8-10m global DGs large use their that the dence the of fact, knowledge future In DGs our near properties. of improve the sample substantially in large can LG, data a the HST for in though CMDs Even et high-quality therein). Gallart 1999; provided references al. and et 1999, Caputo Rich 1999; & al. al. Mighell e.g., et (see Buonanno stel- Hubble 1996; the LG the distant assessing of by for content crucial collected lar were data (HST) while Telescope 1998), Space al. 1994; al. et et Marconi (Smecker-Hane DGs not-too-distant (van for CMDs LG the to VDB). bound hereinafter be not 1999b, may galaxies Bergh of clustering den group Antlia-Sextans by nearest the enriched the obser- that further was The evidence DGs 1998). the nearby Grebel on WIH; Irwin, scenario hereinafter vational (Whiting, 1997, un- galaxies Hau brightness for as & surface also such low but systems very satellites stellar the the their unresolvable of with distant, interaction new M31 the derstanding shed of and formation and to the their Galaxy step on of fundamental only a not a properties light consequence is evolutionary a component(s) the As stellar of evolution. environmen- galactic analysis the the detailed investigate dwarfs on (Grebel to and effects us DGs galaxies allows tal the isolated it fairly of subgroups, since both in Moreover, contains 1999a). total LG Bergh the den the van on 1998; Mateo luminos- turn 1998; global in and the and on composition, ity, depend chemical as history the such star-formation content, parameters the matter fundamental as dark how 1996) estimate the Zijlstra to as & well dif- (Minniti between any, are morphological-types if DGs relationship, ferent nearby the of understand properties galax- to global low-luminosity crucial the on particular, to In questions laboratory ies. open unique a several are par- and address they are formation since (DGs) interesting galaxies galaxy dwarf ticularly both the play them understanding galaxies Among for (LG) evolution. role Group key Local of a properties global The Introduction 1. rudbsddt rvddsvrlde n accurate and deep several provided data based Ground 2 .Cassisi, S. 1 .Buonanno, R. 2 ⋆ n .Nonino M. and ASTROPHYSICS ASTRONOMY 1.11.2018 AND 3 , 4 2 Piersimoni et al.: The Antlia Dwarf Galaxy instruments particularly useful for investigating the stel- IIIaJ plates covering the entire southern sky, who also sug- lar content in the LG galaxies. In this paper, we present gested that this galaxy is probably gravitationally bound the results of an investigation on the Antlia dwarf galaxy to the dwarf NGC3109. After its redis- based on B,V,I data collected with FORS I on VLT dur- cover the global properties of this galaxy were investigated ing the SV program. These photometric data are a plain by Aparicio et al. (1997, hereinafter ADGMD) and by evidence of the VLT capability to investigate DGs in the Sarajedini, Claver & Ostheimer (1997, hereinafter SCO). LG (but see also Tolstoy 1999). These investigations brought out the following character- The layout of this paper is the following: in the next istics: a) low-mass, metal-poor stars are the main stellar section we briefly present the observations and describe component of the galaxy; b) evidence of an age gradient the procedures adopted for the reduction and the calibra- within the galaxy with the young stellar component lo- tion of data. In §3 we discuss the main features of the cated close to the center; c) no evidence of an ongoing CMDs, together with the distance and metallicity esti- process. mates, while the Antlia-Sextans grouping is addressed in A peculiar feature of Antlia is the amount of gas still §4. Finally, a brief summary and the conclusions are out- present in this galaxy, and indeed WIH have inferred a 5 lined in §5. total H I mass of 8 × 10 M⊙. Following the classification suggested by Da Costa (1997), which is based on the ratio between total mass of gas and B integrated luminosity, 2. Observations and data reduction Antlia should be classified as a (dusty) dwarf irregular The data for Antlia have been requested and retrieved (dIrr) rather than as a dwarf spheroidal (dSph) galaxy, electronically from the ESO archive in Garching. The since according to SCO it should also contain interstel- galaxy was observed through the standard Bessell B,V,I lar dust. Oddly enough, it also shows a smooth elliptical filters during the VLT-UT1 SV Program in January 1999 morphology and a very low stellar concentration in the in- using the FORS I camera which covers a 6.8 × 6.8 arcmin nermost regions. These features are quite similar to other field of view at 0.02 arcsec per pixel resolution. The seeing isolated dSph galaxies in the LG such as the Tucana dwarf. was excellent (0.45 - 0.75 arcsec). We used the standard re- However, Tucana does not show a significant amount of duction procedure reported on the Data Reduction Notes gas and therefore Mateo (1998) classified Antlia as a tran- listed in the VLT web pages and the Daophot II package sitional galaxy (dIrr/dSph) together with LGS3, Phoenix, (Stetson, Davis, & Crabtree 1990, and references therein). DDO210, and Pegasus. To improve the detection limit we coadded all the frames Panels a) and b) of Figure 1 show the CMD of Antlia taken with the same filter and then we selected the I coad- in the (I, V −I) and in the (I,B −I) CMD respectively. In ded frame (5400 sec) to create a master catalogue of stellar these diagrams were only plotted stars (≈ 1700) located objects. The stars identified in this search were used as a within 2.5 arcmin from the center of the galaxy. We se- template to fit both the B and the V coadded frames. lected the stars whose centroids in B, V, and I frames were By adopting a SHARP parameter of ±0.5, we detected matched in one pixel (0.2 arcsec) and which satisfy tight in the coadded frames 3711 (B), 2958 (V), and 4583 (I) constraints on photometric accuracy, namely σI ≤ 0.2, stars down to B ≈ 27.0 mag, V ≈ 25.7 mag, and I ≈ 25.5 σB−I ≤ 0.3, σV −I ≤ 0.3, and SHARP parameter ≤±0.5. mag respectively. The calibration was derived by using Figure 1 discloses several interesting features. It is a standard field observed during the same night and by noteworthy the well-populated RGB, extending from I ≈ adopting the average extinction coefficients. Completeness 24 to I ≈ 21.5 mag, and the sizable number of stars tests were performed by randomly adding in each magni- brighter than the TRGB. SCO identified the latter ob- tude bin (0.2 mag) 15-20% of the original number of stars jects as a not-negligible population of Asymptotic Giant to the coadded frames. Only stars that were detected in Branch stars, progeny of an intermediate-age population. the same position and within a magnitude bin of ±0.1 On the basis of stellar counts in nearby fields (ADGMD), mag were considered as recovered. The simulations we per- and of the evidence that these bright stars are not partic- formed suggest that a completeness of the order of 50% ularly concentrated toward the center of the galaxy -they was reached at I ≈ 24.0 mag and B ≈ 25.1 mag respec- appear at distances larger than 1 arcmin- the possibility tively. that they are actually foreground objects cannot be ruled out. 3. The Color-Magnitude diagram By accounting for the RGB intrinsic width, SCO sug- gested that Antlia could contain a sizable amount of inter- The Antlia dwarf galaxy was originally noted by Corwin, stellar dust. This suggestion was mainly based on the slope de Vaucouleurs & de Vaucouleurs (1985) and by both of the RGB upper portion which in their (I, V − I) CMD Feitzinger & Galinski (1985) and Arp & Madore (1987) mimic the slope of objects affected by increasing redden- who also suggested that this stellar system could be a ing. This feature seems not confirmed by present photome- nearby galaxy. This finding was subsequently confirmed try. The (I,B−I) CMD is particularly compelling since in by Fouqu´eet al. (1990) who found, in a detailed HI sur- this plane the slope of the reddening vector is steeper than vey of southern late-type galaxies, that Antlia has a small in the (I, V − I) CMD (see also ADGMD). Even though −1 radial velocity (Vr = 361±2kms ). However, a firm iden- the origin of such a discrepancy cannot firmly established, tification of Antlia was only recently provided by WIH in we suggest that the photometric error (the FWHM of SCO a systematic search for VLSB galaxies in 894 ESO-SRC data is roughly a factor of two larger than in our data) and Piersimoni et al.: The Antlia Dwarf Galaxy 3

Fig. 1. Panel a): The (I,V − I) CMD of Antlia. Panel b): Same as panel a) but for the (I, B − I) CMD. The two redder RGB loci refers to stellar population with an age of 14 (solid line) and 0.8 Gyr (dashed line) respectively. The two isochrones plotted in the blue region refer to stellar ages of 150 (solid line) and 80 Myr (dashed line) respectively. Theoretical prescriptions were plotted in the observative plane by assuming a distance modulus of (m − M)0 = 25.89 mag and a metallicity equal to [M/H]= −1.3. the contamination with foreground objects might have in- therefore that in this galaxy should be present classical troduced a spurious trend (see §4.2 in SCO). Cepheids with periods of the order of few days. The sample of stars located in the blue region of the CMDs -(V − I) < 0.7, (B − I) < 1- plotted in Figure 1 is 3.1. Distance and metallicity. quite interesting. This feature already found by SCO and When dealing with composite stellar population systems by ADGMD suggest the presence of a young stellar pop- for which only the bright end of the CMD is well sampled, ulation. Moreover, their radial distribution clearly shows as in the case of Antlia, the TRGB method turns out to be that this sample is strongly concentrated in the innermost a valuable distance indicator. In fact, this standard candle regions of the galaxy. Therefore we confirm the difference works for all morphological types of galaxies as long as an in the radial distribution between the young and the old old stellar population is present. Moreover the absolute I- stellar component found by SCO and by ADGMD. The Cousins magnitude of the TRGB presents a negligible de- presence of this young stellar component together with pendence on metal content over a wide metallicity range, the evidence of sizable amount of gas are the most clear at least for [M/H] < −0.5. After the first semi-empirical indications that Antlia should be classified as a dIrr rather calibration by Lee, Freedman & Madore (1993, hereinafter than as dSph galaxy. LFM), Salaris & Cassisi (1997, 1998, hereinafter SC97 and In order to supply an estimate of the age of the blue SC98) provided a new theoretical calibration of the TRGB stars we plotted in the CMDs of Figure 1 evolutionary method, characterized by a magnitude shift of ≈ 0.15 - prescriptions (Cassisi 1999) for H and He-burning phases toward brighter magnitudes- in the absolute I magnitude at two different stellar ages, namely t = 80 Myr and t = of the tip. Note that Antlia distance estimates available in 150 Myr. At the same time, we also plotted the location the literature are based on the TRGB calibration provided of the RGB for two old stellar populations at t ≈ 14Gyr by LFM, while we here adopt the SC98 calibration. and t ≈ 0.8Gyr respectively. Theoretical predictions were To estimate the of the TRGB transformed into the observational plane by adopting the we use the differential luminosity function (LF) of RGB bolometric corrections and the color-temperature relations stars. We have not performed any correction since as provided by Green (1988). The adopted distance modu- clearly shown by ADGMD the stars located close to the lus, metallicity and reddening are discussed in the next tip are only marginally affected by foreground contami- section. The comparison between theory and observations nation. The top panel of Figure 2 shows the differential clearly shows that the position of bright blue stars in LF evaluated by adopting a magnitude bin of 0.20 mag. Antlia is finely reproduced by an isochron with an age The error bars for each bin were estimated taking into ranging from 100 to 150 Myr. This age range implies that account both the statistical fluctuations and the correc- the TO of blue stars range from 3.5 to 4.5 M⊙ and tions for completeness. In order to obtain a robust de- 4 Piersimoni et al.: The Antlia Dwarf Galaxy termination of the RGB tip discontinuity, the LF was significantly smaller (ITRGB,0 = 21.66 ± 0.10 mag). This convolved with an edge-detecting Sobel filter [−1, 0, +1]. notwithstanding, our estimate of ITRGB,0 is in good agree- This function shows a sharp peak at ITRGB = 21.7 ± 0.10 ment with the values suggested by SCO (ITRGB,0 = 21.57 mag which marks the appearance of the RGB tip. The mag) and by ADGMD (ITRGB,0 = 21.57 ± 0.05 mag). error was estimated on the basis of the adopted magni- The TRGB method (see LFM, SC97 and SC98) is tude bin. This value is, within current observational un- an iterative procedure which simultaneously gives both certainties, in good agreement with the values obtained the distance and the mean metallicity of the old stel- by SCO (ITRGB = 21.63 ± 0.05) and ADGMD (ITRGB = lar population in the galaxy. The metallicity evaluations 21.64 ± 0.04). Thus supporting a posteriori the negligi- are based on the calibration of the dereddened color in- ble effect of foreground contamination on the LF of RGB dex (V − I) of the RGB half a magnitude below the tip, stars. Our determination is ≈ 0.2 mag fainter than the (V −I)−3.5,0, as a function of the metal content. By adopt- estimate provided by WIH, i.e. ITRGB = 21.4 ± 0.1. This ing the SC98 calibrations we find a distance modulus of discrepancy was already noted by ADGMD who suggested (m − M)0 = 25.89 ± 0.10 mag, i.e. D = 1.51 ± 0.07 Mpc that it could be due to crowding problems at the limiting and (V − I)−3.5,0=1.36 mag, which in turn translates into magnitude in the WIH’s photometry. a mean metallicity of [M/H] ≈−1.3 ± 0.15. This distance modulus is roughly 13% larger than the values found both by SCO ((m − M)0 = 25.62 ± 0.12 mag) and by ADGMD ((m−M)0 = 25.6±0.1 mag). As far as the mean metallic- ity is concerned, our estimate is ≈ 0.3 dex higher than the ADGMD evaluation ([M/H] ≈−1.6 ± 0.1) and ≈ 0.6 dex higher than SCO determination ([M/H] ≈−1.9 ± 0.13). As expected the disagreement is mainly due to the calibration of the TRGB method and of the RGB color index vs. metallicity provided by LFM and by SC98. A thorough discussion of the differences between these cali- brations was already provided by SC98. However, the dis- crepancy with the metallicity evaluation provided by SCO is mainly due to the different approach adopted by these authors for estimating the metallicity along the RGB. In passing we note that our metallicity estimate is in fair agreement with the metallicity obtained by comparing the RGB loci of Antlia directly with the RGB loci of galac- tic globular clusters provided by Da Costa & Armandroff (1990).

3.2. Some hints on RGB stars. The observed color width of the Antlia RGB at I ≈ 22.1 mag -i.e. ≈ 0.5 mag below the RGB tip- is ∆(V −I) ≈ 0.25 mag. However, since at I ≈ 22.1 mag our mean photomet- ric error is σV −I =0.05 mag, it turns out that the intrinsic color width of the RGB is roughly equal to 0.2 mag. If we assume that this color dispersion is due to a spread in Fig. 2. Logarithmic differential LF of RGB stars (solid line) metallicity, the metal content of Antlia should be in the in Antlia. The arrow marks the position of the bin where the range −1.8 < [M/H] < −1.0. However, stellar models TRGB discontinuity was detected. A theoretical LF (dashed supply important information to figure out this problem. line) for an age of 14 Gyr is also shown by adopting a distance In fact, at fixed metal content an increase in age moves modulus equal to (m − M)0 = 25.89 mag. The theoretical LF the RGB loci toward redder colors, while a decrease in was normalized to the observed one, at I ≈ 21.9 mag. Bottom age causes a decrease in the TRGB luminosity (see e.g. panel: The cumulative LF of RGB stars. The dashed lines show SC98; Caputo et al. 1999). Figure 1 shows the theoreti- the two different slopes of the LF. cal prescriptions at fixed metallicity -[M/H] = −1.3- for the RGB loci of an old (14 Gyr, solid line) and a young (0.8 Gyr, dashed line) stellar population. As a result, the By interpolating the maps of Burstein & Heiles (1982) RGB color dispersion in Antlia could be due to a mix of an we estimated E(B − V )=0.03 ± 0.02, which implies a old and of a young/intermediate-age stellar population. foreground reddening of E(V − I)=0.04 ± 0.03. By using The LF shows a further interesting feature: for magni- the extinction relation provided by Cardelli et al. (1989), tudes dimmer than I ≈ 22.3 mag one can notice a signifi- this reddening implies an extinction AI =0.04±0.03 mag. cant increase in the number of RGB stars. This evidence As a consequence, the uncertainty on the I magnitude of is supported by the substantial change in the slope of the the RGB tip is dominated by photometric and complete- cumulative LF plotted in the bottom panel of Figure 2, ness errors since the reddening correction for this band is and can also be identified in the CMDs plotted in Fig- Piersimoni et al.: The Antlia Dwarf Galaxy 5 ure 1. To assess the nature of this feature, Figure 2 shows and TRGB method). The main advantage of our distance the comparison between the observed differential LF and determinations is that they are based on the same stan- the theoretical LF for an age of 14 Gyr. We find that for dard candle and on the same TRGB calibration. However, magnitudes fainter than I ≈ 22.4 mag, the observed stel- distance determinations based both on the LFM and on lar counts are, at the 2σ level, larger than the theoretical the SC98 calibration agree within their errors with the counts expected for an old stellar population. distance scale based on Cepheid Period-Luminosity (PL) This finding could be interpreted as the evidence of relation. This problem might be resolved by a detailed a secondary sample of RGB stars connected with a stel- comparison of distance determinations based on Cepheid lar population younger than the main RGB stellar com- PL relations which account for the metallicity dependence ponent. In fact, for stellar ages lower than ≈ 1Gyr, the and on TRGB method (Bono, Marconi, & Stellingwerf TRGB luminosity becomes quite sensitive to the age. To 1999). constrain the age of such a population we take into account In order to disentangle this thorny problem we esti- the I magnitude difference between the RGB tip associ- mated the distance of NGC3109 by adopting the sample ated with the oldest component, located at ITRGB = 21.65 of classical Cepheids observed in this galaxy by Musella, mag, and the LF discontinuity located at fainter magni- Piotto, & Capaccioli (1998) and the theoretical PL and tudes i.e. I ≈ 22.4mag. By adopting this approach and I PLV relations for Z=0.004 provided by Bono et al. (1999). by assuming for the younger stellar component, the same Interestingly enough, we find that the reddening corrected metallicity of the oldest one, we estimate that its age is distance moduli in these two bands are 25.8±0.1 mag and ≈ 0.7 Gyr. 25.82 ± 0.08 mag respectively. Within the errors, which account only for the intrinsic dispersion, these distance 4. Some hints on the Antlia-Sextans grouping. determinations seem to support more the SC98 than the LFM calibration of the TRGB method. In fact, SC98 de- One of the main reason why the global properties of Antlia rived a distance modulus for NGC3109 of 25.69 ± 0.14 are so interesting is because it should be located beyond mag which is in good agreement with the Cepheid dis- the zero-velocity surface of the LG, and therefore its mo- tance, while Lee (1993) by adopting the LFM calibration tion can be used for providing independent estimates of found a distance modulus of 25.45±0.15 mag and the cor- both the LG dark matter halo and the age of the Uni- responding distance is 15% smaller than the Cepheid dis- verse (Lynden-Bell 1981). However, the location of this tance. On the basis of this finding, it goes without saying galaxy within the LG is still controversial. In fact, SCO that DGs which host both young and old stellar popula- suggested on the basis of its position in the heliocentric tions can play a key role to settle down the dependence radial velocity versus apex angle diagram that it is lo- of Cepheid distance scale on metallicity since these stellar cated near the outer edge of the LG. At the same time, it systems are characterized by a much smaller metallicity was pointed out by ADGMD on the basis of the relative gradient when compared with large spiral galaxies (Ma- velocity between Antlia and NGC3109 that it is unlikely teo 1998). that this pair of galaxies is gravitationally bound. On the other hand, Yahil, Tammann, & Sandage (1977), Lynden- By taking into account the TRGB distances of NGC3109 and Antlia and their separation on the sky Bell & Lin (1977) and, more recently, VDB in a detailed ◦ analysis of the nearest group of galaxies brought out that (≈ 1 .18) we estimated that the projected distance be- Antlia together with Sextans A/B and NGC3109 forms a tween these galaxies is ≈ 140 kpc. The projected distance small cluster of galaxies which is not bounded to the LG, decreases to ≈ 70 kpc if we use the new NGC3109 distance and therefore that it is expanding with the Hubble flow. based on Cepheid distance. By adopting these two differ- He also suggested that Antlia is probably a satellite of ent projected distances and the equation (4) given by VDB we find that this system should have a total mass larger NGC3109 and that this pair to be gravitationally stable 10 10 should contain a sizable amount of dark matter. than 7.8 × 10 M⊙ (TRGB distances) and 4.0 × 10 M⊙ By adopting the distance moduli with the relative er- (Cepheid distance to NGC3109) to be bound. These values rors estimated by SC98 (see their Table 2, column 9 and translates, by assuming mV,0(Antlia)=15.58 (ADGMD) Table 3, column 2) by means of the TRGB method for the and mV,0(NGC3109) = 9.63 (Carignan 1985; Minniti et other three members of this group we find that the cor- al. 1999), into total mass-to-light ratios of (M/LV )0 ≥ 350 responding distances are: D(Sextans A)=1.51 ± 0.1 Mpc, and 170 in solar units. Taken at face value these ratios D()=1.45±0.09 Mpc, and D(NGC3109)=1.37± imply that this system should contain an amount of dark 0.09 Mpc. Taken at face values these distances together matter which is at least a factor of 2-4 larger than in any with the Antlia distance suggest that within the errors other dwarf in the LG (see Table 4 in Mateo 1998). There- these four galaxies are probably located at the same dis- fore, it is seems unlikely that these two galaxies are gravi- tance, thus supporting the finding by VDB that they tationally bound. Finally, we mention that the increase in form a small nearby clustering. Note that with the excep- the mean distances supplies a straightforward support to tion of NGC3109 these distance estimates are systemati- the evidence brought out by VDB that the Ant-Sex clus- cally larger than those adopted by VDB. The discrepancy tering is located beyond the zero-velocity surface of the ranges from 5% for Sextans A up to 13% for Antlia. The LG. reason for the disagreement is partially due to the differ- ence in the TRGB calibration and partially to the differ- ent standard candles adopted by VDB (classical Cepheids 6 Piersimoni et al.: The Antlia Dwarf Galaxy

5. Summary and conclusions. References

Two CMDs -(I, V − I) and (I,B − I)- together with the Aparicio, A., Dalcanton, J.J., Gallart, C. & Martinez-Delgado, LF in the I band, based on photometric data collected D. 1997, AJ, 114, 1447 (ADGMD) with FORS I during the SV program of the VLT were Arp, H. C. & Madore, B. F. 1987, Catalogue of Southern Pe- used for constraining the global properties of Antlia. The culiar Galaxies and Associations (Cambridge: Cambridge new data confirm, as originally suggested by SCO and Univ. Press) ADGMD, that low-mass, metal-poor stars with an age of Bono, G., Marconi, M., & Stellingwerf, R. F. 1999, ApJS, 122, the order of 10 Gyr are the main stellar component of this 167 galaxy and that young blue stars are located close to the Bono, G., Caputo, F., Castellani, V., & Marconi, M. 1999, ApJ, center. The presence of interstellar dust suggested by SCO 512, 711 is not confirmed by current photometric data. Buonanno, R., Corsi, C. E., Castellani, M., Marconi, G., Fusi The comparison between theory and observations sug- Pecci, F., & Zinn, R. 1999, AJ, in press, astro-ph/9907073 gests that the young stellar component is characterized Burstein, D. & Heiles, C. 1982, AJ, 87, 1165 by an age ranging from 100 to 150 Myr, and in turn that Caputo, F, Cassisi, S, Castellani, M., Marconi, G. & Santola- in this galaxy should be present classical Cepheids with mazza, P. 1999, AJ, 117, 2199 periods of the order of few days. By adopting the calibra- Cardelli, J.A., Clayton, G.C., & Mathis, J.S. 1989, ApJ, 345, tions of the TRGB method and of the color index V − I 245 vs. metallicity suggested by SC98 we estimated that the Carignan, C. 1985, ApJ, 299, 59 distance modulus of Antlia is (m − M)0 = 25.89 ± 0.10 Cassisi, S. 1999, in preparation mag -i.e. D = 1.51 ± 0.07 Mpc-, while its mean metal- Corwin, H. G. J., De Vaucouleurs, A. & De Vaucouleurs, G. licity is [M/H] = −1.3 ± 0.15. This distance estimate is 1985, Southern Galaxy Catalog (Austin: Univ. Texas Press) 13% larger than the distance determinations provided by Da Costa, G.S. 1997, in Stellar Astrophysics for the Local SCO and ADGMD, while the mean metallicity is 0.3 dex Group: VIII Canary Islands Winter School of Astrophysics, more metal-rich than the value suggested by ADGMD. eds. A. Aparicio, A. Herrero & F. Sanchez, (Cambridge: The disagreement with previous estimates available in the Cambridge University Press), 351 literature is mainly due to systematic differences in the Da Costa, G.S. & Armandroff, T.E. 1990, AJ, 100, 162 calibration of the TRGB method and of the color index Feitzinger, J. V. & Galinski, T. 1985, A&AS, 61, 503 vs. metallicity relation provided by LFM and SC98. Fouqu´e, P., Bottinelli, L., Durand, N., Gouguenheim, L., & Interestingly enough, we find that the differential Paturel, G. 1990, A&AS, 86, 473 LF shows a secondary peak at I ≈ 22.5 mag which Gallart, C. et al. 1999, ApJ, 514, 665 is at 2σ level larger than theoretical predictions. We Grebel, E. K. 1998, in IAU Symp. 192, The Stellar Content of suggest that this feature could be due to a secondary Galaxies, ed. P. Whitelock, R. Cannon (ASP: young/intermediate-age stellar component. By assuming San Francisco), 1 that this sample of RGB stars presents the same metallic- Green, E. M. 1988, in Calibration of Stellar Ages, ed. A.G. ity of the old component we estimated that its age should Davis Philip (Davis Press: Schenectady), 81 be ≈ 0.7 Gyr. This evidence together with the appearance Lee, M.G., Freedman, W.L., & Madore, B.F. 1993, ApJ, 417, of the blue stars suggests that after the initial burst who 553 (LFM) took place approximately 10 Gyr ago this galaxy experi- Lynden-Bell, D. 1981, Observatory, 101, 111 enced two further star formation episodes ≈ 0.7 and ≈ 0.1 Lynden-Bell, D. & Lin, D. N. C. 1977, MNRAS, 181, 37 Gyr ago. Marconi, G., Buonanno, R., Castellani, M., Iannicola, G., Mo- Finally by using the TRGB method we derived in a laro, P., Pasquini, L. & Pulone, L. 1998, A&A, 330, 453 homogeneous context the distances of Sextans A/B, and Mateo, M. 1998, ARAA, 36, 435 NGC3109 we find that these three galaxies together with Mighell, K.J. & Rich, R.M. 1996, AJ, 111, 777 Antlia are located within the errors at the same distance. Minniti, D. & Zijlstra, A. A. 1996, ApJ, 467, L13 Thus supporting the finding by VDB that these galaxies Minniti, D., Zijlstra, A. A., & Alonso, M. V. 1999, AJ, 117, form a small nearby clustering. The new distances also 881 support the evidence that this grouping could be located Musella, I., Piotto, G. & Capaccioli, M. 1998, AJ, 114, 976 beyond the zero-velocity surface of the LG (VDB). Obvi- Salaris, M., & Cassisi, S. 1997, MNRAS, 289, 406 (SC97) ously new observations aimed at detecting both horizontal Salaris, M., & Cassisi, S. 1998, MNRAS, 298, 166 (SC98) branch stars and RR Lyrae stars as well as at detecting Sarajedini, A., Claver, C.F. & Ostheimer, J.C. Jr. 1997, AJ, and measuring classical Cepheids can supply fundamen- 114, 2505 (SCO) tal constraints on the intrinsic distance and on the global Smecker-Hane, T. A., Stetson, P. B., Hesser, J. E. & Lehnert, properties of this intriguing neighborhood. M. D. 1994, AJ, 108, 507 Stetson, P. B., Davis, L. E. & Crabtree, D. R. 1990, in CCDs Acknowledgements. It is a pleasure to thank V. Castellani and in Astronomy, ed. G.H Jacoby (San Francisco: ASP), 289 M. Marconi for many interesting discussions on an early draft Tolstoy, E. 1999, in VLT Opening Symposium, astro- of this paper. We also warmly acknowledge M. Salaris for many ph/9907029 stimulating suggestions on the content of the paper. Detailed van den Bergh, S. 1999a, The Local Group of Galaxies, to and pertinent comments from the referee, Sydney van den appear in The Astronomy and Astrophysics Review, astro- Bergh, have contributed to improving the content of this paper. ph/9908050 Piersimoni et al.: The Antlia Dwarf Galaxy 7 van den Bergh, S. 1999b, ApJ, 517, L99 (VDB) Whiting, A.B., Irwin, M.J. & Hau, G.K.T. 1997, AJ, 114, 996 (WIH) Yahil, A., Tammann, G. A. & Sandage, A. 1977, ApJ, 217, 903

This article was processed by the author using Springer-Verlag a LTEX A&A style file L-AA version 3.