On the nature of a shell of young stars in the outskirts of the Small Magellanic Cloud
David Martinez-Delgado, A. Katherina Vivas, Eva K. Grebel, Carme Gallart, Adriano Pieres, Cameron P. M. Bell, Paul Zivick, Bertrand Lemasle, L. Clifton Johnson, Julio A. Carballo-Bello, Noelia E. D. Noel, Maria-Rosa L. Cioni, Yumi Choi, Gurtina Besla, Judy Schmidt, Dennis Zaritsky, Robert A. Gruendl, Mark Seibert, David Nidever, Laura Monteagudo, Mateo Monelli, Bernhard Hubl, Roeland van der Marel, Fernando J. Ballesteros, Guy Stringfellow, Alistair Walker, Robert Blum, Eric F. Bell, Blair C. Conn, Knut Olsen, Nicolas Martin, You-Hua Chu, Laura Inno, Thomas J. L. Boer, Nitya Kallivayalil, Michele De Leo, Yuri Beletsky, Ricardo R. Munoz
AAstronomy & Astrophysics manuscript no. aa c (cid:13)
ESO 2019July 12, 2019
On the nature of a shell of young stars in the outskirtsof the Small Magellanic Cloud
David Mart´ınez-Delgado , A. Katherina Vivas , Eva K. Grebel , Carme Gallart , , Adriano Pieres , , CameronP. M. Bell , Paul Zivick , Bertrand Lemasle , L. Clifton Johnson , Julio A. Carballo-Bello , , Noelia E. D. No¨el ,Maria-Rosa L. Cioni , Yumi Choi , , Gurtina Besla , Judy Schmidt , Dennis Zaritsky , Robert A. Gruendl , ,Mark Seibert , David Nidever , Laura Monteagudo , , Mateo Monelli , , Bernhard Hubl , Roeland van derMarel , , Fernando J. Ballesteros , Guy Stringfellow , Alistair Walker , Robert Blum , Eric F. Bell , Blair C.Conn , Knut Olsen , Nicolas Martin , , You-Hua Chu , , Laura Inno , Thomas J. L. Boer , Nitya Kallivayalil ,Michele De Leo , Yuri Beletsky , Ricardo R. Mu˜noz (A ffi liations can be found after the references) ABSTRACT
Context.
Understanding the evolutionary history of the Magellanic Clouds requires an in-depth exploration and characterization of the stellarcontent in their outer regions, which ultimately are key to tracing the epochs and nature of past interactions.
Aims.
We present new deep images of a shell-like over-density of stars in the outskirts of the Small Magellanic Cloud (SMC). The shell, alsodetected in photographic plates dating back to the fifties, is located at ∼ . ◦ from the center of the SMC in the north-east direction. Methods.
The structure and stellar content of this feature were studied with multi-band, optical data from the Survey of the MAgellanic StellarHistory (SMASH) carried out with the Dark Energy Camera on the Blanco Telescope at Cerro Tololo Inter-American Observatory. We alsoinvestigate the kinematic of the stars in the shell using the
Gaia Data Release 2.
Results.
The shell is composed of a young population with an age ∼ Myr, with no contribution from an old population. Thus, it is hard toexplain its origin as the remnant of a tidally disrupted stellar system. The spatial distribution of the young main-sequence stars shows a richsub-structure, with a spiral arm-like feature emanating from the main shell and a separated small arc of young stars close to the globular clusterNGC 362. We find that the absolute g-band magnitude of the shell is M g , shell = − . ± . , with a surface brightness of µ g , shell = . ± . mag arcsec − Conclusions.
We have not found any evidence that this feature is of tidal origin or a bright part of a spiral arm-like structure. Instead, we suggestthat the shell formed in a recent star formation event, likely triggered by an interaction with the Large Magellanic Cloud and / or the Milky Way, ∼
150 Myr ago.
Key words.
Galaxies:individual:Small Magellanic Cloud – Local Group –Magellanic Clouds – Galaxies:photometry –Galaxies:structure
1. Introduction
The Magellanic Clouds are the largest satellites of the MilkyWay (MW), and the only irregular galaxies in its immediate sur-roundings. The existence of one or even two such gas-rich, mas-sive satellites close to a Milky-Way-sized halo has been shownto be quite rare (e.g., Busha et al. 2011; Gonz´alez et al. 2013;Rodr´ıguez-Puebla et al. et al. et al. et al. et al. ◦ across the sky (Putman et al. 1998; Nideveret al. 2010). Both tidal and ram pressure stripping origins havebeen suggested for the Stream (see D’Onghia & Fox 2016 for a review). The mass of ionized and atomic Magellanic gas foundoutside of the Clouds exceeds the remaining H i mass in bothClouds combined (Fox et al. 2014) and is of a similar orderof magnitude as the total mass of the Small Magellanic Cloud(SMC).Another prominent feature is the gaseous Magellanic Bridge,which connects the two Clouds. The Bridge also contains stars,specifically an irregularly distributed young stellar populationwith ages of up to several hundred Myr (e.g., Irwin et al. 1985;Skowron et al. 2014), associations, and star clusters (e.g., Bicaet al. 2008). Unequivocal evidence of intermediate-age stars wasalso found in the Magellanic Bridge area (No¨el et al. 2013; No¨elet al. 2015) as well as unambiguous evidence of a tidal strip-ping scenario (Carrera et al. 2017). The density distribution ofthese inter-cloud stars suggests that they belong to the extendedouter regions of the two Clouds (e.g., Skowron et al. 2014;Jacyszyn-Dobrzeniecka et al. 2017; Wagner-Kaiser & Sarajedini2017). This would support a scenario where the Bridge formedin a close encounter between the Clouds some 250 Myr ago(D’Onghia & Fox 2016 and references therein). On the otherhand, Belokurov et al. (2017) report the discovery of a sep-arate, o ff -set old stellar bridge between the Clouds. However,Jacyszyn-Dobrzeniecka et al. (2019) found the existence of this a r X i v : . [ a s t r o - ph . GA ] J u l art´ınez-Delgado et al.: On the nature of a shell of young stars in the outskirts of the Small Magellanic Cloud b a d ec aa fa A BB C g Fig. 1.
The de Vaucouleurs’ photographic study of the Magellanic Clouds revisited with CCD images obtained with Canon telephotolens from ESO La Silla Observatory. (
Panel A):
Full wide-field mosaic of the Clouds obtained with a Canon 50 lens as describedin Besla et al. (2016). The main features described throughout this paper are marked with letters: The Wing (label a ), the outer armB (label b ) and a well-known foreground Galactic cirrus crossing the field (label c ); ( Panel B) : A zoomed CCD image of the SMCobtained in the same run with a Canon 200 lens showing the shell-like feature (label d) in the outer arm, the SMC south-west plume(label g) and the two globular clusters NGC 362 (label e) and 47 Tuc (label f); (
Panel C) : For a comparison, the sketch based on the1950s photographic plate material of the Clouds from de Vaucouleurs & Freeman (1972) already showed all the features detectedin our modern deep CCD imaging. (AAS. Reproduced with permission)old bridge as controversial, based on the analysis of the distribu-tion of RR Lyrae stars in the OGLE data.Moreover, the outer regions of both Clouds show distortions,clumps, arcs, and related overdensities, both in the direction ofthe Bridge and elsewhere (e.g., Nidever et al. 2013; Casetti-Dinescu et al. 2014; Besla et al. 2016; Belokurov & Koposov2016; Mackey et al. 2016; Belokurov et al. 2017; Pieres et al.2017; Subramanian et al. 2017; Carrera et al. 2017; Mackey etal. 2018; Choi et al. 2018a; Choi et al. 2018b). The last close en-counter between the two Clouds some 100 to 300 Myr ago is notonly consistent with the young ages in the Bridge, but also witha peak in the age distribution of young clusters in both Clouds(Glatt et al. 2010) and with the bimodal ages of Cepheids (Ripepiet al. 2017).The SMC has been particularly a ff ected by past interactionsand shows a highly distorted, amorphous structure in its H i dis-tribution (e.g., Stanimirovic et al. 2004) and in its younger pop-ulations, while its old populations are symmetrically and reg- ularly distributed (e.g. Cioni, Habing & Israel 2000; Zaritskyet al. 2000; Haschke et al. 2012; Jacyszyn-Dobrzeniecka et al.2017; Muraveva et al. 2018). The SMC has a large line-of-sightdepth (e.g., Caldwell & Coulson 1986; Mathewson et al. 1988;Crowl et al. 2001; Subramanian & Subramaniam 2012; Haschkeet al. 2012; Nidever et al. 2013) and its old stellar population isvery extended (No¨el & Gallart 2007; Nidever et al. 2011). Fromtheir study of the 3-D structure of the SMC using Cepheid stars,Ripepi et al. (2017) found 25-30 kpc (see also Scowcroft et al.2016; Jacyszyn-Dobrzeniecka et al. 2016). Apart from repeateddisruptive encounters between the Clouds, it has been suggestedthat the complex structure of the SMC may also be due to adwarf-dwarf merger in the distant past (Bekki & Chiba 2008).One important approach towards understanding the evolu-tionary history of the Magellanic Clouds is through deep, multi-colour mapping of the Clouds, especially of their neglected out-skirts, which can still contain clues about the times and nature ofpast interactions. As part of our deep, wide-field imaging survey Fig. 2.
Image of the Small Magellanic Cloud obtained with a Canon EF 200 mm f / . ◦ × . ◦ , with a pixel scale of 6.7 arcsec pixel − .of faint tidal structures around the Magellanic Clouds and otherMilky Way satellites (Besla et al. 2016; Mart´ınez-Delgado et al.in preparation) using telephoto lenses, we have detected a coher-ent shell-like over-density feature embedded in the di ff use ”outerarm B” (see de Vaucouleurs & Freeman 1972) and previouslyvisible in photographic studies of the Clouds dating back to the1950s (e.g. see Figure 12b in de Vaucouleurs & Freeman 1972and Fig. 1 in this work). Albers et al. (1987) also found evidenceof shell-type structures on the north-east side of the SMC basedon the analysis of star counts from scanned photographic plates,suggesting they could be associated to faint, spiral arm structurein this region of the SMC. The shell-like feature was also pre-viously noted as “a distinct linear feature perpendicular to themain elongation of the SMC” in a stellar density map using onlyupper main-sequence stars selected from the Magellanic CloudPhotometric Survey in Zaritsky et al. (2000, their Fig. 2). Anearly color-magnitude diagram of the north-east outer regionsof the SMC by Brueck & Marsoglu (1978) concluded the pres-ence of a young population ( ∼
60 Myr) and the association ofthree star clusters in the outer arm B , that were interpreted asthe evidence of a recent burst of star formation in this SMC re-gion (Bruck 1980). In this paper, we study the structure, stel-lar and gas content, kinematics and possible formation scenar- ios of this over-density by means of its resolved stellar popu-lations as traced by the imaging data obtained in the Survey ofthe MAgellanic Stellar History (SMASH, Nidever et al. 2017)and the proper motions available from the
Gaia
Data Release 2(DR2) (Gaia Collaboration et al. 2018).
2. The Data
In our data, the SMC over-density was first detected in deepimages taken with an equipment consisting of a Canon EF 200mm f / ◦ × ◦ and a pixelscale of 9.27 arcsec pixel − . A set of 58 individual exposuresof 300 sec were taken in a Luminance filter (see, e.g., Fig. 1 inMart´ınez-Delgado et al. 2015), with a total exposure time of 290minutes. This first image is showed in Fig. 1 ( panel B ), with theSMC over-density marked with label d .Standard data reduction procedures for bias subtraction andflat fielding were carried out using the CCDRED package in the Image Reduction and Analysis Facility (IRAF ). A detailedanalysis of these observations is presented in Mart´ınez-Delgadoet al. (in preparation).To ensure that this shell-like feature was not an artifact ora reflection, a confirmation wide-field image of the SMC wastaken with a di ff erent Canon EF 200 mm f / Astrometry.net (Lang et al. 2010).
SMASH (Nidever et al. 2017) was carried out using the DarkEnergy Camera (DECam, Flaugher et al. 2015) on the 4-mBlanco Telescope at Cerro Tololo Inter-American Observatory,Chile. The survey, covering 480 deg of the sky, aims to explorethe extended stellar populations of the Clouds and their inter-action history. For our current work we use a small subset ofthe SMASH observations. The region of the observed shell inthe Canon images is well covered by SMASH in Fields 9 (cen-tral coordinates (J2000): α = δ = -70:43:05.51), 14( α = δ = -71:15:32.76), and 15 ( α = δ = -72:49:30.00), which are part of a ring of fields surroundingthe central part of the SMC. Each DECam field covers a FoV of3 deg . Each of these fields was observed in the ugriz bands, withdeep observations of 999s in u , i , and z , and 801s in g and r . Inaddition, for each field we obtained three shorter exposures (60s)with large o ff sets to tie all the chips together photometrically andto allow us to cover some of the gaps between the CCDs. Detailson the data processing and photometry can be found in Nideveret al. (2017).The photometry of the DECam images was carried out usingthe photometry program DAOPHOT (Stetson 1987) as incorpo-rated into IRAF. In order to identify point sources in the photom-etry we imposed the following cuts in the DAOPHOT parame-ters: − . < SHARP < . χ < .
0. We also required thatthe Sextractor (Bertin & Arnouts 1996) stellar probability indexwas PROB > .
8. Since the Fields 9, 14, and 15 cover a contigu-ous part of the sky and overlap with each other, we took care toavoid double entries in the final combined catalog. This leavesus with 2,651,378 stars.
3. A shell in the outer region of the SMC
The images of the SMC obtained with two di ff erent Canon 200lenses (Fig.1B and Fig. 2) revealed a clear shell-like feature sit-uated at 1 . ◦ from the centre of the galaxy, just above the barof the SMC (Figure 2). The shape of this over-density is similarto those of the tidal features recently reported in the outskirts ofthe northern LMC (Mackey et al. 2016; Besla et al. 2016) andthe SMC (Pieres et al. 2017) . In this section, we explore the IRAF is distributed by the National Optical AstronomyObservatories, which are operated by the Association of Universitiesfor Research in Astronomy, Inc., under cooperative agreement with theNational Science Foundation. Unfortunately, the position of this shell in the SMC halo (situatedat 8 deg north from the SMC) is outside of the sky area covered in ourphotographic survey, including the Canon 50 images described in Beslaet al. (2016). spatial extent, structure, and stellar populations by analysing thedeep photometry obtained in the SMASH survey.
Figure 3 shows a g versus ( g − i ) color-magnitude diagram(CMD) of our final catalog. The features of the SMC are clearlyseen in this rich diagram, and it shows the wide range of agesof the stellar populations that are typical for this kind of dwarfirregular galaxy. The main sequence (MS) is very extended, trac-ing the SMC’s young populations. The sub-giant branch (SGB)and a wide red giant branch (RGB) are produced by older popu-lations with typical ages of 1 Gyr or more. The vertical red clumpand the red clump (RC) at 19 . (cid:46) g (cid:46)
20 are additional obviousfeatures in the diagram, which trace intermediate-age popula-tions. For a detailed exploration of the star formation history ofthe SMC from wide-field, multi-colour, ground-based imagingwe refer the reader to Rubele et al. (2018) and from very deepHST imaging to Cignoni et al. (2012, 2013).In a first approximation, we selected stars in the main fea-tures described above, which are labeled in Figure 3 as boxes1 (upper MS), 2 (RC) and 3 (lower part of the RGB). Since allof these features are located in the brighter part of the CMD,the spatial coverage is uniform. This is not necessarily true inthe lower part of the CMD since the deepest SMASH exposureswere not dithered and hence the gaps among CCDs would benoticeable.To investigate the nature of the potential shell, we con-structed density maps of the region based on stars in di ff erentparts of the CMD, as shown in Figure 4. In the left panel weshow a map containing stars in the RC and RGB of the SMC( ∼ ,
000 stars from boxes 2 and 3 of Figure 3). The distri-bution of these stars, which trace the intermediate-age and oldpopulations of the SMC, is quite uniform in the region, showingonly a smoothly increasing density toward the centre of the SMClocated in the lower right part of the panel, at α J = δ J = − E ( B − V ) (cid:39) .
05 mag in the northeasternpart of the SMC. The lack of an over-density in the upper panelof Figure 4 is furthermore consistent with Zaritsky et al. (2000)who also found no evidence of the northeastern over-density intheir stellar density maps obtained using giants and red clumpstars as tracers (see their Fig. 3).The region has a very di ff erent appearance when plottingonly young stars belonging to the upper MS of the SMC pop-ulation, as selected by Box 1 in Figure 3. In this map, made with ∼ ,
000 stars, the shell is clearly visible (labeled a in the rightpanel of Fig 4) and a rich structure is associated with it. Besidesthe main shell, we find a spiral-arm-like feature of young starsattached to the shell (labeled b in Fig 4) and a separated small arcsituated ∼ (cid:48) West from the globular cluster NGC 362 (labeled c in the right panel of Fig 4). This last feature has no counter-part in the old population maps, so we conclude that it is not partof a tidal tail from this cluster (see also Carballo-Bello 2019).Instead, it is very interesting that two of the young open clustersdiscussed in Sec. 3.3 (see Table 1) are embedded in the Southern ¡ : ¡ : ¡ : : : : : : ¡ i151617181920212223242526 g - 1 . 5 - 1 . 0 - 0 . 5 0 0.5 1.0 1.5 2.0 2.5g - i1 51 61 71 81 92 02 12 22 32 42 52 6 g Fig. 3.
Left panel:
Colour-magnitude diagram of ∼ . Right panel:
Density countours of the right panel highlighting the densestparts of the CMD, including the old turn-o ff (OTO) and the RGB. The red boxes isolate other features of the SMC stellar population:(1) Upper main sequence, (2) red clump, and (3) lower part of the red giant branch.extreme of this small structure (see Fig.8). The structures a and b (Figure 4, right panel) are also clearly visible in the GALEXimage plotted in Fig. 5, which traces mainly hot young stars,showing an excellent agreement with the position and morphol-ogy of these features as traced in our stellar density maps.Our CMD selection also includes a known streamer of youngstars into the Bridge at δ ∼ -73.3 deg, which is also evident inthe East side of the SMC in our Canon 200 image in Figure 2.This is also the origin of some negative features imprinted in thisregion in the left panel of Figure 4 at δ ∼ -73.3 deg. This is dueto the incompleteness in these more crowded areas, since old redstars are harder to detect near bright blue star formation regions. Fig. 4 indicates that the shell contains a rich population of youngMS stars with a range of ages, which is much less prominent inthe surrounding areas. To study the characteristics of the stel-lar populations present in the shell, we will use de-reddened g ,( g − i ) CMDs and the corresponding colour functions (CF).Given the spatially variable nature of the foreground Galacticreddening, we use the reddening maps of Schlegel, Finkbeiner,& Davis (1998) and combine these with the revised extinctioncoe ffi cients of Schlafly & Finkbeiner (2011) to de-redden eachstar individually. We adopt the recommended “mean” distancemodulus of 18.96 to the SMC as advocated by de Grijs & Bono(2015). Given the areal coverage of the SMASH survey in conjunc-tion with di ff erent environmental e ff ects in each field across thesurvey – such as distinct SFHs, di ff erential reddening, crowd-ing e ff ects, depth, etc. – we opt to compare the de-reddened g ,( g − i ) CMD and CFs of the shell region against two di ff er-ent “control” fields: ConF1 is located close to the shell (see leftpanel in Fig. 4), and ConF2 is on the opposite side of the SMC(not shown in Fig. 4). We aim to investigate the characteristicsof the stellar content in each field, and identify di ff erences in thestellar populations between them.In Fig. 6 we plot the CMD of the stars in the shell box (up-per left panel) and in ConF1 (upper right panel). Box boxes havethe same area. Isochrones from the BaSTI version 5.0.1 library(Pietrinferni et al. 2004 ) of ages of 30 Myr, 150 Myr, 250 Myr,800 Myr, 2.5 Gyr, 6.0 Gyr (Z = = = ii regions, young stars, and Cepheidsin the SMC (Russell & Dopita 1992; Romaniello et al. 2009;Lemasle et al. 2017). Although the SMASH survey utilizes the4-m Blanco telescope and DECam imager, who filter system isdescribed in Abbott et al. (2018), the photometric calibration re-lies on the use of standard stars in the Sloan Digital Sky Survey(SDSS; see Nidever et al. 2017) and as such we adopt the BaSTIisochrones in the SDSS ugriz system for our comparison.From the two upper panels of Fig. 6 it seems that the shellsample contains a significantly higher number of luminous MSstars than the same region of the CMD in the “control” field http:/albinone.oa-teramo.inaf.it : : : : : : : : : : ¡ ¡ ¡ ¡ D E C ( d e g ) : : : : : : : : : : ¡ ¡ ¡ ¡ D E C ( d e g ) Fig. 4.
Left panel:
Density map of stars in boxes 2 and 3 of Figure 3 corresponding to stars on the RC and RGB. The position ofthe control field ConF1 (see Sec. 3.2) is marked with a red quadrilateral.
Right panel:
Density map of stars in box 1 of Figure 3corresponding to stars on the upper MS. The shell-like featured detected in the Canon 200 images is marked with label a . The mapreveals two new features: a spiral arm-like structure (label b ) and an outer, small over-density of young stars (label c ) in the vicinityof the globular cluster NGC 362.ConF1 sample, particularly in the age range 150-250 Myr. It isinteresting, however, that the CMD of ConF1 contains a largerquantity of stars younger that 150 Myr, which are very scarcein the shell sample: from the comparison with the isochrones itcan be concluded that very little star formation has taken placein the shell during the last 150 Myr. A step in the density of starsin the main sequence can also be observed, both in the shell andin the ConF1 CMD at the approximate position of the 2.5 Gyrisochrone, possibly indicating an enhanced period of star forma-tion at intermediate ages. This enhancement is well-known andcorresponds to a common LMC / SMC burst epoch at about 1.5-3Gyr ago (e.g. see Harris & Zaritsky 2002; Weisz et al. 2013).The larger number of young, bright objects in the shell re-sults in a g -band surface brightness for the shell sample that ismore than 0.5 mag arcsec − brighter than the ConF1 control fieldsample (cf. µ g , shell = . ± .
01 and µ g , CF = . ± . − ). We also notice this increased surface bright-ness in the i -band but with a smaller di ff erence (cf. µ i , shell = . ± .
01 and µ i , cont = . ± .
01 mag arcsec − ). Forthe “mean” SMC distance modulus of 18.96, we determine thatthe absolute g - and i -band magnitudes of the shell sample areM g , shell = − . ± .
02 and M i , shell = − . ± .
02 mag, re-spectively.The lower left panel of the Fig. 6 shows the CMD of asynthetic population computed using the BasTI on-line StellarPopulation Synthesis Program , using solar scaled overshoot-ing models. We have assumed a constant star formation rate(SFR) from 13.5 Gyr to 30 Myr ago (the latter is the youngage limit for the BasTI library), and a simplified chemical en-richment law, approximately consistent with that obtained byCarrera et al. (2008) from Ca II triplet spectroscopy, that is,[Fe / H] = -0.99 ( σ = < / H] = -1.29 ( σ = > β = > ff erent colours for the synthetic stars have been used to http://basti.oa-teramo.inaf.it/BASTI/WEB_TOOLS/synth_pop2/ highlight the position in the CMD of stars with di ff erent ages.The same isochrones as in the observed CMDs were plotted asreference. Finally, the lower right panel of Fig. 6 shows the lu-minosity function of the shell, control fields, and that of the syn-thetic CMDs. These luminosity functions allow us to concludethat the observed CMDs are basically complete down to M i = i = ff erences to the SMC fieldpopulations at similar galactocentric radius (for an introductionto the use of the CF for stellar population analysis, see Gallartet al. 2005; see also No¨el et al. 2007 for an application to studySMC field stellar populations). We compare the shell CF withthe CF of the same control field shown in Fig. 4 (ConF1) andcontrol field ConF2 located at the opposite side of the SMC. Theupper panel of Fig. 7 shows, in black and di ff erent line types,the ( g − i ) observed CFs for the shell and the two control fieldsmentioned above. An absolute magnitude cut of M i = g − i ) (cid:39) − . g − i ) (cid:39) RA (h) D E C ( d e g ) Fig. 5.
An ultraviolet image of the SMC covering a similar field to that shown in Fig. 4. The shell features seen by the contributionof young stars are obvious in this near-ultraviolet image of the SMC taken by the Galaxy Evolution Explorer (GALEX, λ e f f ∼ , − − ∼ ∼ ff erences in thestellar content of the shell and the two control fields. The veryprominent blue maximum in the shell CF has very low countsin both ConF1 and ConF2. The di ff erence in the height of thesecond, intermediate colour maximum in the shell and the con-trol fields is not striking, but a di ff erence is nonetheless evident.Taking into account the information on the lower panel of Fig. 7regarding the ages contributing to each feature on the CF (com-plemented by the isochrone information), we conclude that starformation has been very active in the shell in the last (cid:39) (cid:39) g – i ) [mag]161820222426 i [ m ag ] Shell g – i ) [mag]161820222426 CF1 g – i ) [mag]161820222426 i [ m ag ] Synthetic
30 Myr150 Myr250 Myr800 Myr2.5 Gyr6 Gyr13.5 Gyr (n stars )161820222426 ShellCF1Synthetic0 25 50number of stars per pixel0 . . Fig. 6.
Top row:
De-reddened i vs. g − i Hess diagrams of the shell (left) and control field 1 (ConF1, right). The solid linesare BaSTI isochrones with ages of 30, 150, 800 Myr, 2.5, 6, and 13.5 Gyr shifted assuming a distance modulus of 18.96 mag.The metallicity of the oldest isochrone is Z = = Bottom left: i vs. g − i synthetic CMD computed using the BaSTI stellar population synthetic program webtool for a constant SFH from 13.5 Gyr to 30 Myr ago. The solid lines are the same isochrones as shown in the top row. The horizontaldashed line corresponds to our absolute magnitude cut of M i = Bottom right: i luminosity functions of the shell, ConF1, and the synthetic CMD. The horizontal dashed line again corresponds to M i = Fig. 7.
Upper panel : CFs of the shell and control fields (for M i < Lower panel : the same CFs for the observedfields, together with the CF for three populations in di ff erent ageranges, as labelled (see text for details). Cluster Name RA DEC log(age / yr) Ref.B88 14.233 -70.773 8.10 1HW33 14.346 -70.809 8.10 2B139 17.617 -71.561 8.30 1HW64 17.687 -71.338 8.25 1IC1655 17.971 -71.331 8.30 1IC1660 18.158 -71.761 8.20 1L95 18.687 -71.347 8.30 1NGC458 18.717 -71.550 8.15 3HW73 19.108 -71.326 8.15 1 Table 1.
Young Star Clusters in the SMC shell. RA and DEC arein J2000 system. References: (1) Glatt et al. 2010, (2) Piatti etal. 2014, (3) Alcaino et al. 2003.
A number of young star clusters lie spatially coincident with thenortheastern shell feature, with the most prominent clusters ap-pearing as visible over-densities in the upper MS star count mapin Fig. 4. We identify nine young ( < ∼ . > ff use associations and low significance cata-log entries. This sample features three relatively massive clusters -2 -1 0 1 2 3g-i24222018161412 i Fig. 8.
Top panel : MS stellar density map showing the loca-tions of nine young ( < Bottom panel : The combined CMD of the nine youngstar clusters, showing their collective compatibility with a singleage of ∼
160 Myr. We overplot a ∼
160 Myr PARSEC isochrone(log(age / yr) = = A V = = ∼ M (cid:12) ; NGC458, IC1655, IC1660), as well as six other lessmassive systems .A striking feature of this young cluster sample is their uni-formity in age. As determined from isochrone comparisons toobserved CMDs (Alcaino et al. 2003; Glatt, Grebel & Koch2010; Piatti 2014), the cluster ages pile up around ∼
160 Myr The association of three clusters with the position of this featurewas previously mentioned by Brueck & Marsoglou (1978). However,only NGC 458 and L90 actually fall on the shell, while HW62 is clearlyo ff set o ff the stellar over-density towards the SMC center. The ages as-signed to the clusters by these authors, based on photographic plates,have been revisited with our modern observations yielding 160 Myr(instead of 20-50 Myr). 9art´ınez-Delgado et al.: On the nature of a shell of young stars in the outskirts of the Small Magellanic Cloud [log(age / yr) = ± / yr) = ∼
160 Myr ago. This epoch broadly agreeswith the SMC-wide peak of cluster ages observed by Glatt et al.(2010) as well as the putative age of the most recent LMC-SMCinteraction.
OGLE IV (Udalsky et al. 2015; Soszy´nski et al. 2016) uncoveredalmost 5,000 classical Cepheids in the SMC and several tens ofthem spatially overlap with the shell feature (see Fig. 9). SinceClassical Cepheids are young supergiants, their presence in quitelarge numbers underlines significant star formation in the lastfew hundred Myr.Cepheid ages can be computed for individual stars usingperiod-age relations derived from population models (for in-stance, Bono et al. 2005). Using these relations, the ages varyfrom ≈
15 to ≈
500 Myr for SMC Cepheids. The age distribu-tion of SMC Cepheids is known to be bimodal, with two peaksat ≈ ≈ ≈
150 Myr (e.g., Inno et al. 2015; Subramanian et al. 2015;Jacyszyn-Dobrzeniecka et al. 2016; Ripepi et al. 2017). The for-mer peak has been associated with star formation triggered bythe most recent interaction between the LMC and the SMC.Cepheids lying in the shell region span a relatively wide agerange that clearly peaks at 100–130 Myr (see Fig. 10). Such agesare in good agreement with current results for the MS sampleand match very well the ages derived for the young clusters inthe vicinity of the shell . Cepheids therefore support a scenariowhere the shell population is dominated by young stars formedduring a recent star formation event, possibly related to the in-teraction between both Magellanic Clouds. Using the
Gaia
DR2 data (Gaia Collaboration et al. 2018),proper motions for thousands of stars in the SMC outer regionhave recently became available. We use this catalog to investi-gate the kinematic signature from the shell-like feature. Fromthe DR2 database, we select stars surrounding the SMC and ap-ply a series of astrometric cuts. We start by applying a paral-lax cut of ω < . gaia technical note GAIA-C3-TN-LU-LL-124-01) of 1.40 and a cut for the color excess of the stars (asdescribed in Lindegren et al. 2018 by Equation C.2). As astro-metric precision has a strong relationship with the magnitudeof the stars, we additionally apply a cut of G <
18. Finally, totrim down potential MW contamination, we cut out an area with Models that include rotation during the Main Sequence phase leadto Cepheid ages increased by 50 to 100%, depending on the period(Anderson et al. 2016) However, it is important to take into account that the raw age distri-bution of Cepheids cannot be directly interpreted as an age distributionof the underlying star formation because it is convolved with both thestellar IMF and the lifetime of the star within the instability strip duringwhich it would be identified as a Cepheid variable.
Fig. 9.
Classical Cepheids in the OGLE database (red dots) over-plotted on the shell region.radius equal to 3 mas yr − around the systemic proper motion(PM) of the SMC, which we will take to be µ α ∗ , , SMC , µ δ ∗ , , SMC = . ± . , − . ± .
030 mas yr − (Helmi, et al. 2018). Thisleaves us our final selection of stars (seen in upper Figure 11),where the shell-like over-density can be clearly seen (marked bythe purple rectangle).The CMD of the shell region obtained from the Gaia data(bottom panel of Figure 11, expanded down to G =
20 for greatercontext) displays clearly the MS, the red supergiants (RSG), theRC, and the RGB features. Combining the requirement of G <
18 and the apparent locations of the stellar features, we cre-ated masks for each feature. The masks were then applied to thefull sample of SMC stars and the resulting spatial distributionswere plotted (upper Fig. 12). Similar to the analysis earlier inthe paper, the RGB had no apparent correlation with the shell-like region, but both the MS and RSG display over-densities inthe location of the shell. Correspondingly, we select these twosequences from the shell region for our kinematic analysis.For this portion of the analysis, we first subtract the systemicPM from the PM of each star in the shell region. As the featureis noticeably extended on the night sky, we also calculate andremove the viewing perspective for each source, as outlined invan der Marel et al. (2002). These residual proper motions areconverted into a Cartesian frame assuming the kinematically-derived center of ( α J , δ J ) = (16.25 ◦ , − ◦ ), using thetransformations from the Gaia Collaboration (Helmi et al. 2018).The position angle (PA, which in the context of this analysis isdefined as the vector angle of the residual proper motions) iscalculated for each source, defined where 0 ◦ points vertically up-wards in the spatial plot in Figure 12 and increases in a counter-clockwise direction. We compile all of the PAs for all sourcesinto a histogram for easier interpretation (seen in lower panelof Figure 12). The PAs appear to peak around 70 − ◦ , whichpoints roughly radially outwards from the center of the SMC.The scatter from 0 ◦ to 360 ◦ is expected as the average residual ison the order of ∼ . − with average errors of compara-ble magnitude, underscoring that the large peak in PA must be areal signal. This coherent radially outward motion of stars in the Fig. 10.
Age distribution for Cepheids located in the vicinity ofthe shell. Ages have been computed with the period-age relationsfor fundamental and first overtone pulsators of Bono et al. (2005)at Z = i and H α emission Using the atomic hydrogen (H i ) emission maps from theGalactic All-Sky Survey (GASS) third release (Kalberla &Haud 2015), we scanned the velocity channels available in thatsurvey (from −
495 km s − out to +
495 km s − ). The H i gasemission within one square degree centered on α = δ = − . ◦ ) reaches a maximum in the channel v =
198 km s − .more specifically, the velocities of the gas within that region areclose to a Gaussian distribution, with the maximum in v = . − and a dispersion σ = . − . This velocity is verysimilar to the mean velocity of the stars measured by Evans &Howarth (2008)(172 km s − ) with a dispersion of 30 km s − . Theleftmost panel of Fig. 13 shows the emission for this specific ve-locity channel in a log scale. The second panel of Fig. 13 is thedistribution of the mean gas temperature bounded by a largerrange of velocities (184-211 km s − ) in a linear scale, which isvery similar to the previous panel. The rightmost panel shows thegas in the respective velocity channel, in a zoomed region closeto the shell discussed in this paper. The young MS stars are over-plotted on the gas density map as white dots. A visual inspection Fig. 11.
Top panel:
Star map made with all the
Gaia
DR2 (GaiaCollaboration et al. 2018) sources with visibility periods used ≥ phot bp rp excess factor ¡ 1.5 in a 400 arcmin ×
400 arcmin box centered in the SMC. For orientation reference,the LMC is located down and to the left of the plot. The pre-viously described shell and new features showed in Fig.4 (rightpanel) are clearly visible in the
Gaia data.
Bottom panel : CMDof an astrometrically selected sample of
Gaia sources with G <
18 and | µ | < − (marked in gray) with all sources se-lected within the shell region (the exact area selected can be seenin the top panel of Figure 12) over-plotted in purple. The masksfor selecting di ff erent CMD features are also over-plotted: mainsequence (MS, blue), red supergiants (RSG, green), red clump(orange) and red giant branch (red).on the last panel suggests a shift between the projected locationof the young stars and the gas distribution.As we do not have constraints to the distance of the H I gascloud in the Fig. 13, our few arguments are similarities betweenthe velocities and shape of the H I gas cloud and the shell-likefeature of young stars in the Fig. 4. Assuming that the ‘Z’ shapedgas cloud was forming stars in a recent past (150-250 Myr ago), Fig. 12.
Top panel:
Spatial plot of our selected
Gaia sample(marked in grey), discussed in Section 3.5, with the main se-quence (MS, blue) and red supergiant (RSG, green) sequencesshown in Figure 11 overplotted. The shell feature can be seentowards the center-left area of the plot in addition to significantsubstructure in the two sequences spread throughout the SMC.This feature has been marked with a purple box, and all starsthat fall within the box are examined below in addition to beingoverplotted in the CMD for the SMC (seen in the bottom panel ofFigure 11).
Bottom panel:
Combined histogram of the residualproper motion vector angles of the MS and RSG populations af-ter removing the systemic motion of the SMC and correcting forviewing perspective. Only stars that fall within the marked boxin the top of Figure 12 are displayed, indicated by the same col-ors as above. The angle measurement is defined so that a resid-ual vector pointing vertically in the spatial plot corresponds toan angle of 0 ◦ , and the angle increases in a counter- clockwisedirection. A clear preference for a mean residual vector angle of ∼ − ◦ can be seen, which roughly points radially outwardsfrom the center of the SMC.the velocities of the young stars and of the gas cloud could pro-vide insights on what is the dominant process (tidal stripping orram pressure) for decoupling the kinematics of the stars and thegas, at least in that region. However, if the gas cloud formed theshell-like feature, it is not clear what mechanism triggered suchan intense star formation episode ∼
200 Myr ago and why it isno longer acting to form stars. An evidence of the quiescenceof the gas in forming stars is the lack of MS stars between thecurrent position of the gas cloud and the position of the shell fea- ture. Nevertheless, we must take into account the possibility oflarge distances in the line-of-sight between the gas feature andthe young stars, being two completely separate substructures.A comparison of the stellar populations map to the H α map(Fig. 14) from the Magellanic Clouds Emission Line Survey(MCELS; Winkler, Rathore, & Smith 1999) shows a system offilaments that roughly forms a shell with a radius of 0 . ◦ cen-tered at α J = δ J = − α imagesare compared to the H i ATCA + Parkes observations, there ap-pear to be faint (3-5 σ significance) H i emission patches coin-cident with the H α filaments with velocities between ∼ i emission are consistent withthe gas (and driving stellar population) being part of the SMCmain body and not the material being drawn out by the interac-tion with the LMC. Based on the size, 750 pc (0.75 ◦ at 60 kpc)and assuming this interstellar structure arose from the collectivestellar energy feedback from a population originating 150 Myrago, we expect an expansion velocity of 5 km s − which is lowbut not unreasonable compared to other Magellanic supergiantshells (15 to 30 km s − , Book et al. 2008).
4. Discussion
Our analysis of the resolved stellar populations of the elongatedover-density detected in our Canon 200 images reveals that itis the brighter optical part of a more extended structure mainlytraced by blue, young stars distributed in an intricate and com-plex structure. The structure shows a further, outer arc-like fea-ture observed in the projected proximity of the Galactic globu-lar cluster NGC 362 (see bottom panel of Figure 4). The over-density also contains nine young star clusters with ages tightlyclustered at 175 Myr (see Table 1).We do not detect any counterpart of this over-density in thedistribution of the older stars in this area of the SMC. This dom-inant young age of our SMC over-density and the lack of anold stellar remnant in the stellar density map plotted in Fig. 4(left panel) suggest that this feature is neither of tidal originnor the stellar remnant of a tidally disrupted, lower-mass sys-tem as observed in some other dwarf galaxies, e.g., in NGC 4449(Mart´ınez-Delgado et al. 2012) or in Andromeda II (Amorisco etal. 2014). While recently a large number of new, ultra-faint dwarfgalaxies were discovered near the Magellanic System (Drlica-Wagner et al. 2015, 2016), these systems are exclusively de-void of gas and contain very old stellar populations. The exis-tence of a former, more gas-rich small dwarf irregular galaxyis not excluded, but we would expect to also see a clear over-density in the old stellar population map if such a system wereto merge. With the exception of tidal dwarf galaxies, there areno nearby dwarf galaxies known that do not contain old popula-tions (Grebel & Gallagher 2004). Thus, the nature of this over-density seems to be di ff erent from the one discovered at 8 ◦ northof the SMC by Pieres et al. (2017), which is mainly composed ofintermediate-age stars and without significant H i gas. Although − . − . . . . . . . α (h) − − − − − δ ( ◦ ) − . − . . . . . . . α (h) − − − − − δ ( ◦ ) . . . . . . . α (h) − . − . − . − . − . − . − . − − − − − Temperature (K) 10 20 30 40Temperature (K) 10 20 30 40Temperature (K)
Fig. 13. (Left panel):
GASS H i emission for the velocity channel v =
198 km s − , in a log scale. The center of the SMC is markedby a white plus sign. The grey rectangle in this and the following panel is zoomed in the right panel. (Central panel): Same regionas the previous panel, but showing in a linear scale the mean temperature for the mean ± σ distribution in the velocity channel. Aclear ‘Z’ shape is seen with a gradient in temperature towards the center of the SMC. (Right panel): Enlarged region (grey rectanglesin the previous panels) showing the distribution of H i in a linear scale over-plotted to the young main sequence stars as white dots(box 1 in Figure 3). . . . . . . . . . α (h) − . − . − . − . − . − . − . − . − . δ ( ◦ ) − e r g s / c m / s Fig. 14.
Left panel : MCELS H α density map (red, following the color-bar density in the right side) compared to the young stardensity map plotted in Figure 4. Blue dots are the mean stellar position (and its dispersion as error bars) for stars located on the eastof blue dashed line (on the right panel), grouped in 5 bins of declination ( − > δ − Right:
MCELS H α image zoomed in theregion of the shell over-plotted to the GALEX image (see Fig. 5). It shows that a H α filament clearly overlaps the position of thespiral-arm-like feature (labelled b in Fig.4)morphologically similar in appearance to the LMC substructuredescribed by Mackey et al. (2016), the young age of the SMCfeature distinguishes it from that older and much larger LMCover-density. Our stellar density maps of the shell region (Figure 4 andFigure 11) and the GALEX data (Figure 5) confirm with higherresolution the asymmetric structure of young stars originallyfound by Zaritsky et al. (2000) in the SMC outskirts, which con-trasts with the smooth distribution of the older stellar popula- tions. This lack of substructure in the older stellar populationsled these authors to conclude that the dominant physical mech-anism in determining the current appearance of the SMC mustbe recent star formation possibly triggered by a hydrodynamicinteraction of the SMC with a second gas-rich object (e.g., a hotMilky Way halo or the outer gaseous envelope of the LMC). Thecontrast between the small-scale spatial structures in the youngpopulations and the smooth distribution of the intermediate-ageand old populations are even starker in these new data. This find-ing strengthens our argument (see also Zaritsky et al. 2000) thata purely tidal origin for the feature, which would have a ff ectedstars of all ages similarly, is unlikely. Instead, the origin of thefeature is most likely hydrodynamical in nature. Given the co-herence of the feature, the absence of an old stellar remnant, andthe lack of evidence for a strong shock in the H i distribution, wedisfavour models involving a recent accretion event.The elongated shape, estimated absolute magnitude (Sec.3.2), and position of the over-density in the outer region of themain body of the SMC could also suggest a tidal dwarf galaxyas another possible formation scenario for the shell-like struc-ture. The stars in such objects are formed from gas strippedin past encounters (Elmegreen, Kaufman & Thomasson 1993)and / or consist of stars that originally formed in the more mas-sive galaxies participating in the interaction (e.g., Duc 2012).Key requirements for tidal dwarfs include that they are madefrom recycled material, are gravitationally bound, and have de-coupled from their former parent (Duc 2012). They may be ableto survive for at least 3 Gyr in spite of their lack of dark matter(Ploeckinger et al. 2014). In our case, the over-density appears tobe connected with the SMC and there is no evidence to suggestthat it is a separate entity. Hence we also discount the possibilityof a tidal dwarf galaxy.A conspicuous feature in the stellar density map is a spiral-arm-like feature emanating from the over-density (labelled b inFigure 4). Because this stellar arm is only seen in young stars, itis unlikely that it was produced via tidal e ff ects. Instead, the arcmay consist of young stars formed as a result of a low pitch anglespiral density wave in the outer gas. Such features are common,but challenging to detect, in disks well beyond the optical radius(Ferguson et al. 1998; Herbert-Fort et al. 2012). The presenceof a disk population would imply that the young stellar popula-tions of the SMC must have some bulk rotation. However, initial Gaia
DR2-based proper motions and rotation measurements ofthe SMC do not support significant rotation for the young SMCpopulation (van der Marel & Sahlmann 2016), suggesting that,if the density wave scenario is correct, then the observed o ff set isdominated by the progression of the pattern rather than motionof material through it. In addition, the hypothesis of a densitywave also imply an expected o ff set between current (H α ) andpast star formation. We find not detectable o ff set between theH α emission and GALEX sources (see Fig. 14) which, given thetimescales these tracers are sensitive to (10 vs. 10 yrs), placesan upper limit on the pattern speed. Within the new Gaia
DR2catalog, we find among the bright stars the same shell featurein the young stellar populations. Examining this subset of SMCstars reveals coherent motion among the stars but in an outwardradial direction. Given the direction of motion, it continues toappear unlikely that the SMC possesses stellar rotation, at leastrotation that lays primarily in the plane of the sky as with theneighboring LMC. Though, as the vast majority of these starsdo not possess radial velocities, we are unable to fully assess ifthis motion may be due to rotation with a large inclination. Amore holistic analysis of the SMC, including searching for other kinematic substructure, will be required to improve constraintson the degree of stellar rotation.Our analysis of the stellar content of the shell (see Sec. 3.2)indicates that the main di ff erence between the stellar populationof the shell and the surrounding control fields is a recent periodof enhanced star formation in the last (cid:39) ∼
150 Myr, as indicated by the clusters and Cepheids age distri-bution. The comparison of the CFs of the shell and the controlfields also provides hints of a comparatively enhanced star for-mation rate in the total stellar content of this region at interme-diate ages. If we compare these hints on the SFH of the shell andthe control fields with the SFHs of the fields analyzed by No¨elet al. (2009), it can be seen that the SFH of the shell field maypresent similarities with the SFHs of fields located in the Wingarea of the SMC such as qj0112 , qj0111 , and qj0116 , while theSFH of the two control fields may be alike to that of the remain-ing fields analyzed by No¨el et al. (2009), where SFH has beenvery low in the last (cid:39) (cid:39) < <
200 Myr old, that metal-poor gas was injected > g − i ) ∼ ∼
150 Myr ago; or ii) the shell is a region that hadenhanced episodes of star formation at di ff erent epochs, with themost striking one happening ∼
150 Myr ago.It is, thus, interesting to consider the young age of ourSMC substructure in the context of the putative recent inter-action with the LMC about 100 to 300 Myr ago, which leftits signature in the young stellar populations in the MagellanicBridge (e.g., Skowron et al. 2014) and in the age distributionof young populous star clusters in both Clouds (e.g., Glatt et al.2010). Recently, Zivick et al. (2018) used mostly
Hubble SpaceTelescope proper motion data to show that the LMC and SMChave had a head-on collision in the recent past, and it is able toconstrain both the timescale and the impact parameter of this col-lision. Based on their measured proper motions and consideringthe allowed range of masses for the LMC and the Milky Way,they foud that in 97% of all the considered cases, the Cloudsexperienced a direct collision with each other 147 ±
33 Myrago, with a mean impact parameter of 7.5 ± Gaia
DR2 proper motions along the Bridge (Zivick et al. 2019). Theage of our feature falls into this age range as well. Moreover,the three-dimensional structure of young populations traced byCepheids (with ages of about 15–500 Myr) shows a highly asym-metric distribution throughout the SMC: the distribution of clas-sical Cepheids is elongated over 15–20 kpc (e.g., Scowcroft et al.2016; Jacyszyn-Dobrzeniecka et al. 2016), the NE region being younger, closer to the Sun than the SW region (e.g., Haschke etal. 2012; Subramanian et al. 2015; Ripepi et al. 2017). Haschkeet al. (2012) speculate that the displacement and compressionof the gas of the SMC through tidal and ram pressure e ff ectscaused by the interaction between the Magellanic Clouds (andwith the Milky Way) may have locally enhanced star forma-tion, creating some of the irregular, asymmetric features such asthe over-density described in the current paper. Among others,Inno et al.(2015) and Jacyszyn-Dobrzeniecka et al.(2016) alsopropose that the concomitance of the extensive Cepheid forma-tion in the LMC ≈
140 Myr ago and of the younger episode ofCepheid formation in the SMC may be related to the interactionbetween the Clouds. Ripepi et al. (2017) elaborate on this sce-nario and suggest that the relatively young Cepheids ( <
140 Myr)that dominate in the NE region have formed after the dynamicalinteraction that created the Bridge, from gas already shifted bythe interaction.Unfortunately, there is insu ffi cient resolution in any existingnumerical simulations of the LMC-SMC interaction to see thefine structure of our stellar density map in Fig. 4. Our resultsclearly motivate more detailed studies of the internal structureand star formation induced in the SMC by the LMC-SMC (andthe MW) interaction. In particular, if the stellar arc is a spiralstructure, these data likely disfavor models where the SMC isoriginally modeled as a non-rotating spheroid. This result fur-ther highlights the discrepant kinematics and spatial distributionof the SMC younger and older stellar populations, the origin ofwhich is currently unknown. Acknowledgements.
We thank David Hogg for his help with the astrometry so-lution of the image used in this work. DMD thanks Prof. Ken Freeman for adiscussion about the detection of the SMC shell in the photographic plates of theClouds taken in the 1950s. DMD also thanks the hospitality and fruitful discus-sion about this work with the European Southern Observatory Garching head-quarters sta ff during his stay as part of the ESO visitor program in September2017. DMD, EKG, BL and LI acknowledge support by Sonderforschungsbereich(SFB) 881 “The Milky Way System” of the German Research Foundation(DFG), particularly through sub-projects A2, A3 and A5. DMD acknowledgessupport from the Spanish MINECO grant AYA2016-81065-C2-2. This projectused data obtained with the Dark Energy Camera (DECam), which was con-structed by the Dark Energy Survey (DES) collaboration. M-RC and CB ac-knowledge support from the European Research Council (ERC) under theEuropean Union (cid:48) s Horizon 2020 research and innovation program (grant agree-ment No. 682115). BCC acknowledges the support of the Australian ResearchCouncil through Discovery project DP150100862. T.d.B. acknowledges supportfrom the European Research Council (ERC StG-335936). This work has beensupported by the Spanish Ministry of Economy and Competitiveness (MINECO)under grant AYA2014-56795-P. MS acknowledges support from the ADAP grantNNX14AF81G. R.R.M. acknowledges partial support from project BASALAFB-170002 as well as FONDECYT project N ◦ Gaia ( ), processed bythe Gaia
Data Processing and Analysis Consortium (DPAC, ). Funding for the DPC hasbeen provided by national institutions, in particular the institutions participatingin the
Gaia
Multilateral Agreement.
References
Abbott, T. M. C., Abdalla, F. B., Allam, S., et al. 2018, ApJS, 239, 18Albers, H., MacGillivray, H. T., Beard, S. M., & Chromey, F. R. 1987, A&A,182, L8Alcaino, G., Alvarado, F., Borissova, J., & Kurtev, R. 2003, A&A, 400, 917Amorisco, N. C., Evans, N. W., & van de Ven, G. 2014, Nature, 507, 335Anderson, R. I., Saio, H., Ekstr¨om, S., Georgy, C., Meynet, G, 2016, A&A 591,A8Bekki, K., & Chiba, M. 2008, ApJ, 679, L89Bekki, K. 2011, MNRAS, 416, 2359Belokurov, V., & Koposov, S. E. 2016, MNRAS, 456, 602Belokurov, V., Erkal, D., Deason, A. J., et al. 2017, MNRAS, 466, 4711Bertin, E., & Arnouts, S. 1996, A&AS, 117, 393Besla, G., Kallivayalil, N., Hernquist, L., et al. 2007, ApJ, 668, 949Besla, G., Kallivayalil, N., Hernquist, L., et al. 2010, ApJ, 721, L97Besla, G., Mart´ınez-Delgado, D., van der Marel, R. P., et al. 2016, ApJ, 825, 20Bica, E., Bonatto, C., Dutra, C. M., & Santos, J. F. C. 2008, MNRAS, 389, 678Bono, G., Marconi, M., Cassisi, S., Caputo, F., Gieren, W., Pietrzy´nski, G., 2005ApJ 621, 966Book, L. G., Chu, Y.-H., & Gruendl, R. A. 2008, ApJS, 175, 165Boylan-Kolchin, M., Besla, G., & Hernquist, L. 2011, MNRAS, 414, 1560Bruck, M. T. 1980, A&A, 87, 92Brueck, M. T., & Marsoglu, A. 1978, A&A, 68, 193Busha, M. T., Marshall, P. J., Wechsler, R. H., Klypin, A., & Primack, J. 2011,ApJ, 743, 40Caldwell, J. A. R., & Coulson, I. M. 1986, MNRAS, 218, 223Carballo-Bello, J. A. 2019, MNRAS, 486, 1667Carrera, R., Gallart, C., Aparicio, A., et al. 2008, AJ, 136, 1039Carrera, R., Conn, B. C., No¨el, N. E. D., Read, J. I., & L´opez S´anchez, ´A. R.2017, MNRAS, 471, 4571Casetti-Dinescu, D. I., Moni Bidin, C., Girard, T. M., et al. 2014, ApJ, 784, L37Choi, Y., Nidever, D. L., Olsen, K., et al. 2018, ApJ, 866, 90Choi, Y., Nidever, D. L., Olsen, K., et al. 2018, ApJ, 869, 125Cignoni, M., Cole, A. A., Tosi, M., et al. 2012, ApJ, 754, 130Cignoni, M., Cole, A. A., Tosi, M., et al. 2013, ApJ, 775, 83Cioni, M.-R. L., Habing, H. J., & Israel, F. P. 2000, A&A, 358, L9Crowl, H. H., Sarajedini, A., Piatti, A. E., et al. 2001, AJ, 122, 220de Grijs, R., & Bono, G. 2015, AJ, 149, 179de Vaucouleurs, G., & Freeman, K. C. 1972, Vistas in Astronomy, 14, 163D ´Onghia, E., Besla, G., Cox, T. J., Hernquist, L., 2009, Nature, 460, 605D’Onghia, E., & Fox, A. J. 2016, ARA&A, 54, 363Drlica-Wagner, A., Bechtol, K., Ryko ff , E. S., et al. 2015, ApJ, 813, 109Drlica-Wagner, A., Bechtol, K., Allam, S., et al. 2016, ApJ, 833, L5Duc, P.-A. 2012, Astrophysics and Space Science Proceedings, 28, 305Ederoclite, A., & Cepa, J. 2010, Astronomical Data Analysis Software andSystems XIX, 434, 253Elmegreen, B. G., Kaufman, M., & Thomasson, M. 1993, ApJ, 412, 90Evans, C. J., & Howarth, I. D. 2008, MNRAS, 386, 826Fox, A. J., Wakker, B. P., Barger, K. A., et al. 2014, ApJ, 787, 147Gaia Collaboration, Brown, A. G. A., Vallenari, A., et al. 2018, A&A, 616, A1Gaia Collaboration, Helmi, A., van Leeuwen, F., et al. 2018, A&A, 616, A12Gallart, C., Zoccali, M., & Aparicio, A. 2005, ARA&A, 43, 387Glatt, K., Grebel, E. K., & Koch, A. 2010, A&A, 517, A50Gonz´alez, R. E., Kravtsov, A. V., & Gnedin, N. Y. 2013, ApJ, 770, 96Grebel, E. K., & Gallagher, J. S., III 2004, ApJ, 610, L89Harris, J., & Zaritsky, D. 2002, Modes of Star Formation and the Origin of FieldPopulations, 285, 313Haschke, R., Grebel, E. K., Du ff au, S., 2012, AJ 144, 107Herbert-Fort, S., Zaritsky, D., Moustakas, J., et al. 2012, ApJ, 754, 110Inno, L., Bono, G., Romaniello, M., Matsunaga, N., Pietrinferni, A., Genovali,K., Lemasle, B., Marconi, M., Primas, F., 2015 ASPC 491, 265Irwin, M. J., Kunkel, W. E., & Demers, S. 1985, Nature, 318, 160Jacobs, B. A., Rizzi, L., Tully, R. B., et al. 2009, AJ, 138, 332Jacyszyn-Dobrzeniecka, A. M., Skowron, D. M., Mr´oz, P., Skowron, J.,Soszy´nski, I., Udalski, A., Pietrukowicz, P., Kozlowski, S., Wyrzykowski,L., Poleski, R., Pawlak, M., Szyma´nski, M. K., Ulaczyk, K., 2016, AcA 66,149Jacyszyn-Dobrzeniecka, A. M., Skowron, D. M., Mr´oz, P., et al. 2017, ActaAstron., 67, 1Jacyszyn-Dobrzeniecka, A. M., Soszy´nski, I., Udalski, A., et al. 2019,arXiv:1904.08220Kalberla, P. M. W., & Haud, U. 2015, A&A, 578, A78 Kallivayalil, N., van der Marel, R. P., & Alcock, C. 2006, ApJ, 652, 1213Kallivayalil, N., van der Marel, R. P., Alcock, C., et al. 2006, ApJ, 638, 772Kallivayalil, N., van der Marel, R. P., Besla, G., Anderson, J., & Alcock, C. 2013,ApJ, 764, 161Kroupa, P., Tout, C. A., & Gilmore, G. 1993, MNRAS, 262, 545Lang, D., Hogg, D. W., Mierle, K., Blanton, M., & Roweis, S. 2010, AJ, 139,1782Lemasle, B., Groenewegen, M. A. T., Grebel, E. K., et al. 2017, A&A, 608, A85Mackey, A. D., Koposov, S. E., Erkal, D., et al. 2016, MNRAS, 459, 239Mackey, D., Koposov, S., Da Costa, G., et al. 2018, ApJ, 858, L21Mart´ınez-Delgado, D., Romanowsky, A. J., Gabany, R. J., et al. 2012, ApJ, 748,L24Mart´ınez-Delgado, D., D’Onghia, E., Chonis, T. S., et al. 2015, AJ, 150, 116Mathewson, D. S., Cleary, M. N., & Murray, J. D. 1974, ApJ, 190, 291Mathewson, D. S., Ford, V. L., & Visvanathan, N. 1988, ApJ, 333, 617Meaburn, J. 1980, MNRAS, 192, 365Muraveva, T., Subramanian, S., Clementini, G., et al. 2018, MNRAS, 473, 3131Nidever, D. L., Majewski, S. R., Butler Burton, W., & Nigra, L. 2010, ApJ, 723,1618Nidever, D. L., Majewski, S. R., Mu˜noz, R. R., et al. 2011, ApJ, 733, L10Nidever, D. L., Monachesi, A., Bell, E. F., et al. 2013, ApJ, 779, 145Nidever, D. L., Olsen, K., Walker, A. R., et al. 2017, AJ, 154, 199No¨el, N. E. D., & Gallart, C. 2007, ApJ, 665, L23No¨el, N. E. D., Aparicio, A., Gallart, C., et al. 2009, ApJ, 705, 1260No¨el, N. E. D., Conn, B. C., Carrera, R., et al. 2013, ApJ, 768, 109No¨el, N. E. D., Conn, B. C., Read, J. I., et al. 2015, MNRAS, 452, 4222Pasquali, A., Larsen, S., Ferreras, I., et al. 2005, AJ, 129, 148Piatek, S., Pryor, C., & Olszewski, E. W. 2008, AJ, 135, 1024Piatti, A. E. 2014, MNRAS, 445, 2302Pieres, A., Santiago, B. X., Drlica-Wagner, A., et al. 2017, MNRAS, 468, 1349Pietrinferni, A., Cassisi, S., Salaris, M., & Castelli, F. 2004, ApJ, 612, 168Putman, M. E., Gibson, B. K., Staveley-Smith, L., et al. 1998, Nature, 394, 752Ripepi, V., Cioni, M.-R. L., Moretti, M. I., Marconi, M., Bekki, K., Clementini,G., de Grijs, R., Emerson, J., Groenewegen, M. A. T., Ivanov, V. D.,Molinaro, R., Muraveva, T., Oliveira, J. M., Piatti, A. E., Subramanian, S.,van Loon, J. Th., 2017, MNRAS 472, 808Rodr´ıguez-Puebla, A., Avila-Reese, V., & Drory, N. 2013, ApJ, 773, 172Romaniello, M., Primas, F., Mottini, M., Pedicelli, S., Lemasle, B., Bono, G.,Franc¸ois, P., Groenewegen, M. A. T., Laney, C. D., 2008, A&A 488, 731Rubele, S., Girardi, L., Kerber, L., et al. 2015, MNRAS, 449, 639Rubele, S., Pastorelli, G., Girardi, L., et al. 2018, MNRAS, 478, 5017Russell, S. C., & Dopita, M. A. 1992, ApJ, 384, 508Schlegel, D. J., Finkbeiner, D. P., & Davis, M. 1998, ApJ, 500, 525Scowcroft, V., Freedman, W., L., Madore, B. F., Monson, A. J., Persson, S. E.,Rich, J., Seibert, M., Rigby, J. R., 2016 ApJ 816, 49SSchlafly, E. F., & Finkbeiner, D. P. 2011, ApJ, 737, 103Skowron, D. M., Jacyszyn, A. M., Udalski, A., et al. 2014, ApJ, 795, 108Soszy´nski, I., Udalski, A., Szyma´nski, M. K., Skowron, D., Pietrzy´nski,G., Poleski, R., Pietrukowicz, P., Skowron, J., Mr´oz, P., Kozlowski, S.,Wyrzykowski, L., Ulaczyk, K., Pawlak, M., 2016, AcA 65, 297Stanimirovic, S., Staveley-Smith, L., Dickey, J. M., Sault, R. J., & Snowden,S. L. 1999, MNRAS, 302, 417Stanimirovi´c, S., Staveley-Smith, L., & Jones, P. A. 2004, ApJ, 604, 176Stetson, P.B. 1987, PASP, 99, 191Subramanian, S., & Subramaniam, A. 2012, ApJ, 744, 128Subramanian, S., Subramaniam, A., 2015, A&A 573, A135Subramanian, S., Rubele, S., Sun, N.-C., et al. 2017, MNRAS, 467, 2980Udalski, A., Szymanski, M. K., Szymanski, G., 2015, AcA 65, 1van der Marel, R. P., Alves, D. R., Hardy, E., & Suntze ff , N. B. 2002, AJ, 124,2639van der Marel, R. P., & Sahlmann, J. 2016, ApJ, 832, L23Weisz, D. R., Dolphin, A. E., Skillman, E. D., et al. 2013, MNRAS, 431, 364Winkler, P. F., Rathore, Y., & Smith, R. C. 1999, New Views of the MagellanicClouds, 190, 97Zaritsky, D., Harris, J., Grebel, E. K., & Thompson, I. B. 2000, ApJ, 534, L53Zivick, P., Kallivayalil, N., van der Marel, R. P., et al. 2018, ApJ, 864, 55Zivick, P., Kallivayalil, N., Besla, G., et al. 2019, ApJ, 874, 78 Astronomisches Rechen-Institut, Zentrum f¨ur Astronomie derUniversit¨at Heidelberg, M¨onchhofstr. 12–14, 69120 Heidelberg,Germany Cerro Tololo Inter-American Observatory, National OpticalAstronomy Observatory, Casilla 603, La Serena, Chile Instituto de Astrof´ısica de Canarias (IAC), Calle V´ıa L´actea s / n,E-38205 La Laguna, Tenerife; Spain Departamento de Astrof´ısica, Universidad de La Laguna (ULL), E-38206 La Laguna, Tenerife; Spain Laborat´orio Interinstitucional de e-Astronomia - LIneA, Rua Gal.Jos´e Cristino 77, Rio de Janeiro, RJ - 20921-400, Brazil Observat´orio Nacional, Rua Gal. Jos´e Cristino 77, Rio de Janeiro, RJ- 20921-400, Brazil Leibniz-Institut f¨ur Astrophysik Potsdam, An der Sternwarte 16,D-1442 Potsdam, Germany Department of Astronomy, University of Virginia, 530 McCormickRoad, Charlottesville, VA 22904, USA CIERA and Department of Physics and Astronomy, NorthwesternUniversity, 2145 Sheridan Road, Evanston, IL 60208, USA Instituto de Astrofisica, Pontificia Universidad Catolica de Chile, Av.Vicuna Mackenna 4860, Macul 7820436, Santiago, Chile Chinese Academy of Sciences South America Center for Astronomy,National Astronomical Observatories, CAS, Beijing 100101, China Department of Physics, University of Surrey, Guildford GU2 7XH,UK Steward Observatory, University of Arizona, 933 North CherryAvenue,Tucson, AZ 85721 Department of Physics, Montana State University, P.O. Box 173840,Bozeman, MT 59717-3840, USA Astrophysics Source Code Library, University of Maryland, 4254Stadium Drive College Park, MD 20742, USA National Center for Supercomputing Applications, 1205 West ClarkSt., Urbana, IL 61801, USA Department of Astronomy, Universityof Illinois, 1002 W. Green Street, Urbana, IL 61801, USA TheObservatories of the Carnegie Institution for Science, 813 SantaBarbara St., Pasadena, CA 91101 National Optical Astronomy Observatory, 950 N. Cherry Ave,Tucson, AZ 85719, USA Astrophoton Observatory, Edtfleck 12, A-4542 Nussbach, Austria Space Telescope Science Institute, 3700 San Martin Drive,Baltimore, MD 21030 Center for Astrophysical Sciences, Department of Physics &Astronomy, Johns Hopkins University, Baltimore, MD 21218, USA Observatori Astron`omic, Universitat de Val`encia, Paterna (Val`encia),Spain Center for Astrophysics and Space Astronomy, Department ofAstrophysical and Planetary Sciences,University of Colorado 389 UCBBoulder Colorado 80309-0389, USA Department of Astronomy, University of Michigan, 1085 S.University Ave., Ann Arbor, MI 48109 USA Research School of Astronomy and Astrophysics, AustralianNational University, Canberra, ACT 2611, Australia Universit´e de Strasbourg, CNRS, Observatoire astronomique deStrasbourg, UMR 7550, F-67000 Strasbourg, France Max-Planck-Institut f¨ur Astronomie, K¨onigstuhl 17, D-69117Heidelberg, Germany Institute of Astronomy and Astrophysics, Academia Sinica, P.O. Box23-141, Taipei 10617, Taiwan, R.O.C. Department of Astronomy, University of Illinois at Urbana-Champaign, 1002 West Green Street, Urbana,IL 61801, USA Las Campanas Observatory, Carnegie Institution of Washington, LaSerena, Chile32