Satellites and central galaxies in SDSS: the influence of interactions on their properties
Valeria Mesa, Sol Alonso, Georgina Coldwell, Diego García Lambas, Jose Luis Nilo Castellón
MMNRAS , 1–14 (2020) Preprint 1 January 2021 Compiled using MNRAS L A TEX style file v3.0
Satellites and central galaxies in SDSS: the influence of interactionson their properties
Valeria Mesa, , (cid:63) Sol Alonso, Georgina Coldwell, Diego García Lambas and J.L. Nilo Castellon Departamento de Astronomía, Facultad de Ciencias, Universidad de La Serena, Av. Juan Cisternas 1200 Norte, La Serena, Chile Instituto Argentino de Nivología Glaciología y Ciencias Ambientales (IANIGLA-CONICET), Parque Gral San Martín, CC 330, CP 5500, Mendoza, Argentina Departamento de Astronomía y Geofísica, CONICET, Facultad de Ciencias Exactas, Físicas y Naturales (FCEFN)–UNSJ, Av. José Ignacio de la Roza Oeste 590,J5402DCS, San Juan, Argentina Instituto de Astronomía Teórica y Experimental, (IATE-CONICET), Laprida 854, X5000BGR, Córdoba, Argentina
Accepted XXX. Received YYY; in original form ZZZ
ABSTRACT
We use SDSS-DR14 to construct a sample of galaxy systems consisting of a central object andtwo satellites. We adopt projected distance and radial velocity difference criteria and imposean isolation criterion to avoid membership in larger structures. We also classify the interactionbetween the members of each system through a visual inspection of galaxy images, finding ∼ of the systems lack evidence of interactions whilst the remaining ∼ involvesome kind of interaction, as inferred from their observed distorted morphology. We have con-sidered separately, samples of satellites and central galaxies, and each of these samples weretested against suitable control sets to analyse the results. We find that central galaxies show-ing signs of interactions present evidence of enhanced star formation activity and youngerstellar populations. As a counterpart, satellite samples show these galaxies presenting olderstellar populations with a lower star formation rate than the control sample. The observedtrends correlate with the stellar mass content of the galaxies and with the projected distancebetween the members involved in the interaction. The most massive systems are less affectedsince they show no star formation excess, possibly due to their more evolved stage and lessgas available to form new stars. Our results suggest that it is arguable a transfer of materialduring interactions, with satellites acting as donors to the central galaxy. As a consequence ofthe interactions, satellite stellar population ages rapidly and new bursts of star formation mayfrequently occur in the central galaxy. Key words: galaxies: general - galaxies: interactions - galaxies: statistics
Throughout the history of the Universe, galaxy-galaxy interactionsplay a crucial role in galaxy formation and evolution according tothe hierarchical model of structure formation, by linking togetherstar formation processes with galaxies growth (Sol Alonso et al.2006; Woods & Geller 2007; Ellison et al. 2010; Lambas et al.2003, 2012; Mesa et al. 2014). On the observational side, severalworks (e.g. Yee & Ellingson 1995; Kennicutt 1998; Rogers et al.2009; Ellison et al. 2011; Lambas et al. 2012) have shown thatinteractions between galaxies are powerful mechanisms to trigger star formation activity (SFA) . Taking into account statistical anal-ysis of galaxy pairs, Barton et al. (2000) and Lambas et al. (2003)have found that the proximity in radial velocity and projected dis-tance is correlated to an increase of the SFA. Also, Balogh et al.(2004) found a correlation between star formation rate (SFR) and (cid:63)
E-mail: [email protected] environment attributable to starbursts induced by galaxy-galaxy in-teractions.In this scenario, the interaction between galaxies of similarsize and mass would be a very good laboratory to study the conse-quences of the fusion processes in the formation of galaxies. Unfor-tunately, these systems are statistically insignificant, focusing thestudies in systems with higher mass, accompanied by the so-calledsatellite galaxies. While galaxy collisions are expected to be moreviolent, encounters between galaxies with their smaller compan-ions, would be the most common, because low-light galaxies aremore frequent in the Universe. In this line, Daddi et al. (2005) us-ing HST data to study the evolution of early type galaxies, foundsignatures in the B band, compatible with an ongoing merger orcannibalism of satellites. Trujillo et al. (2006, 2007) have studiedthe size evolution of compact massive galaxies, finding that drymerger scenario can be considered as a reasonable mechanism forthe subsequent evolution of these galaxies, since this type of merg-ers are not efficient at forming new stars, but are efficient in increas- © 2020 The Authors a r X i v : . [ a s t r o - ph . GA ] D ec Mesa et al. ing the size of the objects. These results have also been reportedby van Dokkum et al. (2008) using deep and high-resolution im-ages and moreover by van der Wel et al. (2014) through the use ofspectroscopy and multiwavelength photometry from the 3D-HSTsurvey combined with CANDELS imaging. In addition, Bernardi(2009) showed that early-type BCGs identified in the Sloan Digi-tal Sky Survey (SDSS) grew from many dry minor mergers. Morerecently, Vulcani et al. (2014) analysed the relation between colourand structure within galaxies using GAMA survey, showing thatearly type galaxies are associated with multiple collapse and merg-ing events.There are different and varied studies about the impact of thefusion of galaxies in these systems. They show that the presence ofa close companion generates a clear increase in the morphologicalasymmetries of the galaxy even at 50 h − kpc away (Patton et al.2016). Numerical simulations have shown that galaxies grows byaccreting other galaxies, mostly smaller companions (Shao et al.2018; Nipoti et al. 2018; Forbes et al. 2016). Naab et al. (2009)have studied the influence of minor mergers on the evolution ofelliptical galaxies. Their results show that this type of encounterswould be the drivers for the late evolution of sizes and densities ofearly-type galaxies. In this sense, Oser et al. (2010) also providesthat the formation of galaxies has different phases, and that an ex-tended phase in evolution consists of an important growth due tothe accretion of smaller satellite stellar systems. On the other hand,Hirschmann et al. (2015) showed that stellar accretion from minormergers of satellite galaxies results in steep negative metallicity andcolour gradients and slightly positive age gradients, successfullymatching the observed profiles of local galaxies. All this allows usto conclude the importance of the study of satellite galaxy systemsand the analysis of their properties.Stierwalt et al. (2015) presented a systematic study about starformation and the further processing of the interstellar medium inthe interaction between dwarf galaxies. The authors concluded thatthe interactions between dwarf galaxies are important conductorsof galactic evolution in the low mass end, but ultimately the envi-ronment is responsible for the extinction of star formation.For all environments, bulge-dominated galaxies have a colour-magnitude diagram dominated by red galaxies which depends lin-early on absolute magnitude (Hogg et al. 2004; Blanton et al. 2005).Many authors have also studied the relationship of satellite galax-ies with their environment, based on numerical simulations. For in-stance, Barber et al. (2015) predict a statistical excess of satelliteswhose main axis aligns with the direction to the central galaxy. Ev-idence of this was found in the satellite population of M31, whichsuggests that tidal effects may have played an important role inits evolution. Tempel et al. (2015) also noted this alignment, sug-gesting that filaments identified on larger scales can be reflected inthe positions of the satellite galaxies that are very close to its cen-tral galaxies. Sales et al. (2015) examined the colours of satellitegalaxies in the Illustris simulation. They found that the satellitesroughly trace the distribution of dark matter in their system, andthat in massive systems, red satellites dominate and are distributedmore abruptly than the blue population, while for the lower massprimary galaxies, the satellites are mostly blue. Additionally, obser-vational studies suggest that the properties of a satellite galaxy arestrongly correlated with those of its central galaxy (Weinmann et al.2006). Using a large galaxy group catalogue constructed from theSDSS, van den Bosch et al. (2008) have proved that satellites areredder and slightly more concentrated than central galaxies of thesame stellar mass. This scenario points to strangulation as the mainmechanism that operates on satellite galaxies, and that causes their transition from the blue to the red sequence. These results were alsoreported by Wetzel et al. (2012) from SDSS data too.Furthermore, Deason et al. (2014) research the frequency ofmajor mergers between dwarf galaxies in the Local Group usingcosmological simulations. They found that ∼ of dwarf satel-lite galaxies with M ∗ > M (cid:12) inside the virial radius, expe-rienced a major fusion with a ratio of stellar mass close to 0.1from z = 1 , with a lower fraction for dwarf galaxies of smallermass. They found that satellite-satellite mergers also occur withinthe main halo after virial infall, catalysed by the large fraction ofdwarf galaxies that fall down into the group. The fraction of fusionsdoubles for dwarf galaxies outside the virial radius as well that themost distant dwarf galaxies in the local group are the most likely tohave experienced a recent major merger. Tinker & Wetzel (2010)tested how galaxies evolve onto the red sequence, finding that ∼ % of satellite galaxies being red or quenched, involving that ∼ % efficient, andalso showed that all satellites quench on their first infall.More recently, De Lucia et al. (2019) used semi-analytic mod-els to study the time-scales in which star formation is suppressedin satellite galaxies. Finding that environmental processes playa marginal role in passive galaxies with stellar mass larger than M ∗ > M (cid:12) . However, the models need to be improved to pre-dict the behaviour of less massive galaxies, as is the case of satellitegalaxies.There are few observational studies on this topic, and in thosefound in the literature, the relation between satellites and their cen-tral galaxy stands out, but without a deep analysis on the possibleinteraction between satellite galaxies and how these influences theglobal system properties. Gutiérrez et al. (2006) studied a sampleof 31 satellites orbiting isolated giant spiral galaxies finding threecases of clear interactions between the satellites. Four of the galax-ies in their sample are among the objects with higher starformingactivity. In contrast, the only two galaxies of the sample that arenot forming stars are also members of these pairs. They proposethat the presence of the bridges connecting a satellite with theircompanions, and the comparatively large amount of gas are signsof mass transfer from one galaxy to the other. This is probably in-hibiting the star formation activity in the donor and enhancing it inthe accreting galaxy. The stripping suffered by the satellite galax-ies could also be responsible for its morphological changes. Also,Knobel et al. (2013) showed that the fraction of satellite galaxiesthat are red, is systematically higher than that of centrals, and thatthe satellite quenching efficiency (i.e. the probability that a satelliteis quenched because it is a satellite rather than a central) is indepen-dent of stellar mass. These effects are likely to remain even at highredshifts. More recently, Pasquali et al. (2019) examined the phys-ical properties of satellite galaxies in the projected phase-space oftheir host environment, for satellites inside one virial radius of theirhost. They show that low mass satellites are more sensitive to en-vironment and that the general characteristics depend on the timespent in their host environment.This paper is structured as follows: Section 2 describes thedata used in this work. In Section 3 we show the procedure used MNRAS , 1–14 (2020) atellites and central galaxies in SDSS to construct the satellite galaxies catalogue, explaining the classi-fication process of the different types of interactions, and we alsopresent the procedure for building the control samples. An analysisof star formation rates, colours and stellar population, and their dif-ferences with the control samples is described in Section 4. Finallyin Section 5, we summarise the main conclusions.Throughout this paper we adopt a cosmological model char-acterised by the parameters Ω m = 0 . , Ω Λ = 0 . and H =70 km s − Mpc − . This work is based on data provided by the the Sloan Digital SkySurvey (SDSS; York et al. 2000), one of the most successful sur-veys in the history of astronomy. Over years of operations (SDSS-I, 2000-2005; SDSS-II, 2005-2008; SDSS-III, 2008-2014) SDSSdata have been annually released to the scientific community. Thelatest generation of the SDSS data (SDSS-IV, 2014-2020; Blan-ton et al. 2017) is extending precision cosmological measurementsto a critical early phase of cosmic history (eBOSS), expanding itsgalactic spectroscopic survey to the north and south hemispheres(APOGEE-2), using for first time the Sloan spectrograph, perform-ing spatially resolved maps of individual galaxies (MaNGA). Inthe present work we consider spectroscopic data from SDSS DataRelease 14 (Abolfathi et al. 2018, SDSS-DR14). For this sample,k-corrections band-shifted to z = 0 . were calculated using thesoftware k-correct_v4.2 of Blanton & Roweis (2007). Forthe data set, k-corrected absolute magnitudes were calculated fromPetrosian apparent magnitudes converted to the AB system. In ouranalysis we will use the u, g and r-bands in the ugriz system.In this paper we will carry out an analysis about the star for-mation efficiency, colour distributions, age of stellar populationsbased on different parameters such as the D n (4000) spectral in-dex, SF R/M ∗ , all available in the SDSS spectroscopic database.We obtain all data catalogues through SQL queries in CasJobs .From this catalogue we use the star formation rate normalised to thetotal mass in stars estimated from the SDSS fibre, log ( SF R/M ∗ ) ,taken from Brinchmann et al. (2004). As it has been discussed bythese authors, aperture effects could be important for the most mas-sive galaxies. Therefore, for satellite galaxies it is not expected tobe an issue here, but this effect must be taken it into account forcentrals. We have compared the angular size of the SDSS fibre andthe radius containing half of the galaxy light, ( r ). We find that forthe majority of our centrals, the fiber size is within a 50 % fractionof this radius. For this reason, we have compared the derived fi-bre SFR with the values obtained for the global SFR, as calculatedby Brinchmann et al. (2004) finding no significant differences inthe resulting values. So, although for central galaxies the fibre SFRestimate corresponds only to a small central portion, it provides re-liable SFR global determinations. We use the total stellar masses Log ( M ∗ /M (cid:12) ) calculated using the Bayesian methodology, andmodel grids described in Kauffmann et al. (2003). We also use thespectral index D n (4000) , as an indicator of the age of stellar popu-lations. This spectral discontinuity occurring at 4000Å (Kauffmannet al. 2003) arises by an accumulation of a large number of spectrallines in a narrow region of the spectrum, an effect that is importantin the spectra of old stars. We have adopted Balogh et al. (1999)definition of D n (4000) as the ratio of the average flux densities http://skyserver.sdss.org/casjobs/ in the narrow continuum bands (3850-3950 Å and 4000-4100 Å).Finally, to discriminate between bulge and disc-types galaxies, weuse the concentration index C (Abraham et al. 1994), a well testedmorphological classification parameter (Strateva et al. 2001), alsoused as a good stellar-mass tracer ( M ∗ ) and an indirect index ofthe SFR (Deng 2013). Yamauchi et al. (2005) performed a galaxymorphological classification using the C parameter, finding a verygood agreement with the visual classification. Our sample was created based on a combination of photometricand dynamical criteria. First, a base sample composed of abright galaxy plus fainter surrounding sources were defined.Galaxies brighter than M r = − . mag. were selected ascentral galaxies and objects that lies inside r p < h − kpc and ∆ V <
500 km s − , restricted to a difference of 1.5 mag-nitudes fainter regarding to the central galaxy, were classified assatellites of the main object. These criteria were chosen takinginto account our previous experience in the study of galaxy pairs.For instance, Lambas et al. (2003); Sol Alonso et al. (2006)found that r p < h − kpc and ∆ V <
350 km s − wereconvenient thresholds for stellar formation activity induced by theinteractions, and and that this is triggered in values lower than r p < h − kpc and ∆ V <
100 km s − . On this basis, we thenwork with larger samples reaching up to r p < h − kpc and ∆ V <
500 km s − Mesa et al. (2014), and r p < h − kpc and ∆ V <
500 km s − Mesa et al. (2018). Considering now thepresence of another galaxy in the system, we decided to increasethe projected distance r p , taking into account the criteria used byDuplancic et al. (2018) who used the value of r p < h − kpc and ∆ V <
500 km s − to identify small galaxy systems.At this stage, we only selected those systems with twosatellites. In addition, was necessary to define isolation crite-ria in order to ensure that the dynamics of our systems is notdominated by larger virialised structures where they could beimmersed. For instance, an adequate isolation criteria was takeninto account considering that within a radius of 500 h − kpc and ∆ V < − there should not be a brighter galaxythan one magnitude fainter than the central galaxy. O’Mill et al.(2012) adopted similar thresholds to find galaxy triplets and alsoDuplancic et al. (2018), with the aim to define an homogeneousselection criteria of small galaxy systems. Furthermore, thesystems were requested to be within z < Log ( M ∗ ) , r-band M r , concentration index ( C ), D n (4000) and Log ( SF R/M ∗ ) for central galaxies and their satellites.In order to quantify these differences we have computed themean values of Log ( M ∗ ) finding 9.68 ± ± M r we find values of -19.23 ± ± C ) we found values of2.51 ± ± D n (4000) values correspondingto 1.51 ± ± Log ( SF R/M ∗ ) C = r /r is the ratio of Petrosian 90 %- 50% r-band light radiiMNRAS , 1–14 (2020) Mesa et al.
Table 1.
Classification, number of systems and percentages in the differenttypes.Classification Number of systems PercentagesNon Interacting 338 80.67%Interaction between satellites 3 0.72 %Interaction satellites central galaxy 78 18.62%Total 419 100% values of -10.85 ± ± D n (4000) and higher ones of Log ( SF R/M ∗ ) indicates active star formationactivity and younger stellar populations). Also, in Fig 2 we showconcentration index ( C ) vs M r for central galaxies and satelliteshighlighting again the great differences between both samples.Different authors (e.g. Gadotti 2009; Mesa et al. 2014; Morselliet al. 2017) propose that an adequate threshold to separate galaxiesbetween bulge-types and disc-type is 2.5. From this plot we cansee that the satellite sample presents all kinds of morphology,conversely the central galaxies exhibit higher values of C index,indicating that a higher fraction of galaxies in this sample presentsbulge morphology. Once the catalogue of satellite galaxies was obtained, we per-formed an eye-ball classification using the SDSS-DR14 imagingavailable in SkyServer in order to distinguish between differentclasses of interactions. The systems were classified according tothree categories:(i) the main object with two satellites without apparent interactions,(ii) mutual interaction between the satellites, or(iii) between the main object and some of its satellites.This procedure was made with the purpose of analysing therelation between the different components of the systems. Fig. 3shows examples of the different classifications. It is important tonote on how different types of interactions influence on the accre-tion process of the material in primary galaxies. This visual classi-fication is important since it allows to classify different type of in-teractions and besides it permits to detect spurious systems and/ormisclassification from SDSS. We have used this method previouslywith excellent results (e.g. Alonso et al. 2007; Lambas et al. 2012;Mesa et al. 2014). This technique also allows us to clean the sampleof galaxies immersed in groups, undetected by the software due thegalaxies only have photometric information. We found that 94%of the systems were classified into these subsamples. The remain-ing that do not fulfil these three categories were excluded from thepresent analysis. Table 1 provides the classification, number of sys-tems and percentages in this sample of satellite galaxies. https://skyserver.sdss.org/dr14/en/tools/chart/listinfo.aspx Figure 1.
From top to bottom: Distribution of
Log ( M ∗ ) , M r , concen-tration index ( C ), D n (4000) and Log ( SF R/M ∗ ) for central galaxies(dashed lines) and satellites (solid lines). Figure 2.
Concentration index ( C ) vs M r for central galaxies (magentacircles) and satellites (cyan squares). MNRAS000
Concentration index ( C ) vs M r for central galaxies (magentacircles) and satellites (cyan squares). MNRAS000 , 1–14 (2020) atellites and central galaxies in SDSS Figure 3.
Examples of galaxy systems images with different classification: Systems without interaction (left),interaction between satellites (middle) andinteraction with main galaxy (right). The inbox in middle panel shows a zoom to the interacting satellites.
In order to understand the behaviour of our systems, we used acontrol sample for each catalogue, central and satellite galaxies,with the aim to compare different properties with respect toisolated galaxies. Therefore we use a Monte Carlo algorithmwe build control samples of galaxies without a companion bymatching the redshift ( z ), r-band absolute magnitude ( M r ) andconcentration index ( C ) distributions of our samples, following thework of (Perez et al. 2009). The process was done simultaneouslyfor each parameter, randomly matching bins of 0.5 mags for Mr,0.015 for z and 0.25 for C, respectively. These control sampleshave a larger number of galaxies than the main samples allowingto have confident statistical testing sets.We build control samples matching the values of the pa-rameters listed above (Fig. 4). This procedure was performedseparately, both for central galaxies and their satellites. In allcases we performed a Kolmogorov-Smirnov (KS) test and weobtained p > . , hence we can not reject the null hypothesisthat the samples were drawn from the same distribution. Then,any difference in the galaxy properties is associated only with theinteraction, consequently, by comparing the results we estimate thereal difference between satellite or central galaxies (with possibleinteractions) and galaxies without a companion, unveiling theeffect of morphology or luminosity on this features. Through this work, it is intended to have a greater knowledge ofhow the properties of the primary galaxies are affected by the pres-ence of their satellites, as well as the role of the interactions onthem.To this, we will focus on the analysis of the properties of thecentral galaxies of our systems under study. For this purpose, thesystems have been split into two groups, those with the interactionbetween the satellites together with those with signatures of inter-action between the central galaxy and the satellite, with the aim to have better statistics. And on the other hand, we have consideredthose systems without obvious interaction.
In order to characterise the colours of the central galaxies belongingto our sample, in figure 5 the colour-magnitude diagrams ( M u − M r and M g − M r versus M r ) of these galaxies are observed, with andwithout interaction with their satellites. The comparison sample isalso included. We can observe that central galaxies are mostly pop-ulating the so-called "red sequence", while the galaxies with inter-action and those belonging to the control sample, tend to be locatedin the region of the "green valley". This behaviour is also reflectedin the colour index distributions, M u − M r and M g − M r , in thesame figure.The fraction of objects with bluer colours than the medianof the sample has been calculated. Fig 6 shows these fractions asa function of Log ( M ∗ ) , exhibiting that the central galaxies withinteractions have a slightly higher fraction of blue colours withrespect to the other samples. All the uncertainties were derivedthrough a bootstrap resampling technique (Barrow et al. 1984). To study the age of the stellar populations, the spectral index, D n (4000) , and the specific star formation rate, Log ( SF R/M ∗ ) ,will be used. The standard distributions of these parameters foreach type of system are shown in Fig 7, together with their controlsample defined in previous Section. It can be seen that the centralgalaxies of our systems constitute an old population aged and withlow star formation. With some differences only in the case to havesignatures of interactions, the central galaxies of the systems showmore efficient star formation activity and younger stellar popula-tion, with respect to the control samples.To quantify these differences, we have calculated the fractionof young galaxies with stellar formation, that is, the fraction of ob-jects with values of D n (4000) that are below the median of the totalsample, and values of Log ( SF R/M ∗ ) above. This also allows torule out the influence that the masses of galaxies may have.In Fig 8 these fractions are shown as a function of the mass MNRAS , 1–14 (2020)
Mesa et al.
Figure 4.
Left:Distribution of z, M r concentration index ( C ) for central galaxies (dashed lines) and control sample (shaded).Right: Satellites (solid lines) and control sample (shaded). Figure 5.
Left: Colour magnitude diagram for central galaxies in systems without interaction (red squares) and systems with interaction between satellites orwith main galaxy (blue triangles) and control sample (grey dots), and Mu − Mr normalised distribution. Green open circles represents central galaxies insystems with double interactions. Points in the upper left corner represent data errors, for each subsample.Right: Colour magnitude diagram and Mg − Mr normalised distribution. MNRAS000
Left: Colour magnitude diagram for central galaxies in systems without interaction (red squares) and systems with interaction between satellites orwith main galaxy (blue triangles) and control sample (grey dots), and Mu − Mr normalised distribution. Green open circles represents central galaxies insystems with double interactions. Points in the upper left corner represent data errors, for each subsample.Right: Colour magnitude diagram and Mg − Mr normalised distribution. MNRAS000 , 1–14 (2020) atellites and central galaxies in SDSS Log(M*)10.8 11 11.2 11.40.20.40.60.8 Log(M*)10.8 11 11.2 11.40.20.40.60.81
Figure 6.
Fraction of galaxies with M u − M r and M g − M r smaller thanthe median of the sample for central galaxies in systems without interaction(red solid) and systems with interaction between satellites or with maingalaxy (blue dotted) and control sample (shaded). of the galaxies. In this case, there are notable differences in the be-haviour of galaxies with or without interactions with their satellites,the former being the ones with the highest fractions, decreasing asthe mass of the objects increases. The control sample shows an in-termediate behaviour between these two. At greater mass, no dif-ferences are observed between the samples, within the errors con-sidered. This section is devoted to the study of the properties of the satellitegalaxies in our sample, in order to understand how and to what ex-tent they are affected by the processes of interaction between themor with their central galaxy. Similarly to the previous section, theyhave been grouped into two categories: those with some kind ofinteraction and those with no apparent interaction.
With the main goal of characterising the colours of the satellitegalaxies belonging to our sample, in Fig 9 can be seen the colour-magnitude diagrams of the satellite galaxies, with and without in-teractions with their respective central galaxies, including the com-parison sample. In it, it is observed how the galaxies belonging tosystems are found to a greater extent populating the "red sequence"and "green valley", unlike the control galaxies that have a morespread distribution. In addition, the distribution of the Mu − Mr and Mg − Mr colour indexes, which accounts for this behaviour,is presented in the same figure.The fraction of objects with bluer colours than the median ofthe sample was computed for satellite galaxies. Fig 10 display thisfractions as a function of Log ( M ∗ ) , showing that satellite galaxies Figure 7.
Distribution of D n (4000) spectral index and Log ( SF R/M ∗ ) for central galaxies in systems without interaction (red solid) and systemswith interaction between satellites or with main galaxy (blue dotted) andcontrol sample (shaded). Log(M*)10.8 11 11.2 11.40.51 Log(M*)10.8 11 11.2 11.40.20.40.60.811.2
Figure 8.
Fraction of galaxies with
Log ( SF R/M ∗ ) higher and D n (4000) lower than the median of the sample for central galaxies in systems withoutinteraction (red solid) and systems with interaction between satellites orwith main galaxy (blue dotted) and control sample (shaded).MNRAS , 1–14 (2020) Mesa et al. have a lower fraction of blue colours with respect to the comparisonsample.From these analysis we argue for an early enhancement of thestar formation activity in the satellites due to the strong effect ofthe central galaxy, producing a rapid consumption of the gas andtherefore at the present they are redder, with older stellar popula-tion and lower SFR. Hence, the gravitational/tidal interactions de-veloping between satellites and central during the satellites orbitsare the physical mechanism responsible for removing gas from thesatellites, causing a fast quenching of their stellar populations. Thisfact had already been noticed by Gutiérrez et al. (2006) and Knobelet al. (2013).
Fig 11 shows the normalised distributions of the spectral index D n (4000) and the specific star formation rate Log ( SF R/M ∗ ) foreach type of system. The control sample defined in Section 3.2 isalso included. A bimodal distribution can be seen for the galaxiesbelonging to the systems under study. An important presence ofgalaxies with older stellar populations and low stellar formation,regardless of the type of interaction, is also observed. However, un-like the central galaxies, a counterpart with active star formationand younger populations is present. Both samples present similarbehaviours, differentiating themselves from their control sample,composed of a younger population, with clear efficient star forma-tion. This behaviour is evidenced in Fig 12, where fractions ofyoung galaxies with signatures of star formation are calculated.These fractions have been determined according to the medianvalue of the total sample. These values are below the values foundfor the central galaxies, revealing more differences between thesamples, and hence the importance of studying them independently.In this case, it is observed how these amounts increase at a lowerstellar mass. However, there seem to be no differences between thesamples, because regard they interact or not, both are always keptbelow the control sample. In view of these results, we can confirma trend that has already been observed in previous section and bydifferent authors (e.g. van den Bosch et al. 2008; Wetzel et al. 2012;Tinker & Wetzel 2010; Hirschmann et al. 2014; Oman & Hudson2016). Satellite galaxies are quenched with respect to the field ascan be deduced from its comparison with the control sample, andstrangulation would be a possible mechanism for this to occur. Among the systems with evident interactions, there is the particu-lar case of those in which the two satellites are in interaction withtheir central galaxy. In our sample we found seven systems thatpresent this feature. Some examples can be seen in the Fig. 13. Wethink there must have special consideration with these, since theprocesses are likely to be more efficient, and that can be reflectedin their properties. For instance, we have carried out the statisticsof the indicators of star formation and age of stellar populations,finding values in the median of
Log ( SF R/M ∗ ) and D n (4000) of -10.19 ± ± Log ( SF R/M ∗ ) and D n (4000) of -9.80 ± ± As shown in previous sections, we find that interactions can af-fect diverse galaxy properties by inducing different process. Theseproperties may change in different ways depending on whether thegalaxies are centrals or satellites. Besides, they may also correlatewith the relative mass of the interacting galaxies. An important is-sue to explore is the dependence of the effects on projected dis-tance between the interacting galaxies. To further explore this fact,in this section we study the fractions of younger, star-forming andbluer galaxies, according to the median of each sample. This studyconsiders the total sample, and a subsample of galaxies with rela-tive projected distance r p < kpc , corresponding to the medianvalue of the total sample.In Table 2 the percentages of galaxies with values of log ( SF R/M ∗ ) , D n (4000) and ( M u − M r ) below the medianof their corresponding sample are computed, for the total sampleand for the subsample composed by systems at closer projectedseparations ( r p < kpc ). Following the development of thiswork, the values are presented separately for the samples of centralgalaxies and their satellites.In this table the different behaviour of central galaxies andsatellites in response to interactions can be clearly seen. In thecase of centrals, there is always a higher percentage of thosesystems that have interactions, and this is even more evident forthe subsample with smaller values of r p , on the other hand insystems without interactions signs, these values remain nearlyconstant. This is different in the case of satellites, although thereis a difference in systems with interactions within the sample withlower r p , this difference is not as significant as in the previouscase. Systems without interactions maintain a similar proportionaccording to projected distance. By comparison with the totalsample, it can be seen that these fractions are not significantlyaffected by the interactions.On the other hand, we have estimated the minimal enclosingcircle of each system, taking into account the projected distancebetween the three members of the system. An interesting analysisresults from the study of the global properties of the systems de-pending on the radius of this circle. For this aim, we compute thesum of the star formation rates for the three members of a givensystem ( SF R c + SF R + SF R ). Fig 14 shows the behaviourof the total star formation rate as a function of the radius of theminimal enclosing circle ( r mec ) . We must emphasise that this ra-dius expands in different ranges according to the type of interac-tion observed in the system, this result is expected. Also, we plot Log ( SF R total /M ∗ total ) vs ( r mec ) as a capture of the global starformation enhancement recently happened because of interactions.We can see that for systems without interactions these values re-main almost constant, within the errors considered. However, insystems with interactions, the star formation increases for smaller r mec values. In dashed lines, the contribution of the central galaxyhas been added, where it can be seen that it is the one that dominatesthe trend. This is likely due to the fact that since central galaxies are MNRAS000
Log ( SF R/M ∗ ) and D n (4000) of -10.19 ± ± Log ( SF R/M ∗ ) and D n (4000) of -9.80 ± ± As shown in previous sections, we find that interactions can af-fect diverse galaxy properties by inducing different process. Theseproperties may change in different ways depending on whether thegalaxies are centrals or satellites. Besides, they may also correlatewith the relative mass of the interacting galaxies. An important is-sue to explore is the dependence of the effects on projected dis-tance between the interacting galaxies. To further explore this fact,in this section we study the fractions of younger, star-forming andbluer galaxies, according to the median of each sample. This studyconsiders the total sample, and a subsample of galaxies with rela-tive projected distance r p < kpc , corresponding to the medianvalue of the total sample.In Table 2 the percentages of galaxies with values of log ( SF R/M ∗ ) , D n (4000) and ( M u − M r ) below the medianof their corresponding sample are computed, for the total sampleand for the subsample composed by systems at closer projectedseparations ( r p < kpc ). Following the development of thiswork, the values are presented separately for the samples of centralgalaxies and their satellites.In this table the different behaviour of central galaxies andsatellites in response to interactions can be clearly seen. In thecase of centrals, there is always a higher percentage of thosesystems that have interactions, and this is even more evident forthe subsample with smaller values of r p , on the other hand insystems without interactions signs, these values remain nearlyconstant. This is different in the case of satellites, although thereis a difference in systems with interactions within the sample withlower r p , this difference is not as significant as in the previouscase. Systems without interactions maintain a similar proportionaccording to projected distance. By comparison with the totalsample, it can be seen that these fractions are not significantlyaffected by the interactions.On the other hand, we have estimated the minimal enclosingcircle of each system, taking into account the projected distancebetween the three members of the system. An interesting analysisresults from the study of the global properties of the systems de-pending on the radius of this circle. For this aim, we compute thesum of the star formation rates for the three members of a givensystem ( SF R c + SF R + SF R ). Fig 14 shows the behaviourof the total star formation rate as a function of the radius of theminimal enclosing circle ( r mec ) . We must emphasise that this ra-dius expands in different ranges according to the type of interac-tion observed in the system, this result is expected. Also, we plot Log ( SF R total /M ∗ total ) vs ( r mec ) as a capture of the global starformation enhancement recently happened because of interactions.We can see that for systems without interactions these values re-main almost constant, within the errors considered. However, insystems with interactions, the star formation increases for smaller r mec values. In dashed lines, the contribution of the central galaxyhas been added, where it can be seen that it is the one that dominatesthe trend. This is likely due to the fact that since central galaxies are MNRAS000 , 1–14 (2020) atellites and central galaxies in SDSS Figure 9.
Left: Colour magnitude diagram for satellite galaxies in systems without interaction (red squares) and systems with interaction between satellites orwith main galaxy (blue triangles) and control sample (grey dots), and Mu − Mr normalised distribution. Green open circles represents satellite galaxies insystems with double interactions. Points in the upper left corner represent data errors, for each subsample.Right: Colour magnitude diagram and Mg − Mr normalised distribution. Table 2.
Percentages of galaxies with values of log ( SF R/M ∗ ) , D n (4000) and ( M u − M r ) smaller than the median of the sample for central and satellitegalaxies. Systems with r p < h − kpc Systems with r p < h − kpc Ranges Interacting % Non-interacting % Interacting % Non-interacting % Central galaxies log ( SF R/M ∗ ) > − . ± ± ± ± D n (4000) < . ± ± ± ± ( M u − M r ) < . ± ± ± ± log ( SF R/M ∗ ) > − . ± ± ± ± D n (4000) < . ± ± ± ± ( M u − M r ) < . ± ± ± ± more massive than satellites, and in star-forming systems, SFR in-creases with M*.Although not similar work have been done using systems withtwo satellite galaxies, these results are consistent with those foundby Lambas et al. (2003); Alonso et al. (2004); Patton et al. (2013,2020) where the proximity between two paired galaxies triggerstheir star formation activity. In this sense, it is possible speculatethat the tidal effects produced by the central galaxies in both satel-lites could strip them of their gas reservoir, and produce efficientbursts of star formation. Following the previous analysis, in this subsection we study theefficiency of interactions to trigger the formation of stars in thesystem considered as a whole. To achieve this goal, we use the thesum of the star formation rates, previously calculated, and we com-pute the sum of the stellar masses too ( M ∗ c + M ∗ + M ∗ ). Fig. 15shows the behaviour of the total star formation rate as a function ofthe total stellar mass. Also, we show Log ( SF R total /M ∗ total ) as afunction of total stellar mass, analogous to the previous section, asan indicator of recent star formation. It is clearly seen by compari-son of the samples that interacting systems show an enhanced star MNRAS , 1–14 (2020) Mesa et al.
Log(M*)9 9.5 10 10.500.20.40.60.811.2 Log(M*)9 9.5 10 10.500.20.40.60.811.2
Figure 10.
Fraction of galaxies with Mu − Mr and Mg − Mr lower thanthe median of the sample for satellite galaxies in systems without interaction(red solid) and systems with interaction between satellites or with maingalaxy (blue dotted) and control sample (shaded). Figure 11.
Distribution of D n (4000) spectral index and Log ( SF R/M ∗ ) for satellite galaxies in systems without interaction (red solid) and systemswith interaction between satellites or with main galaxy (blue dotted) andcontrol sample (shaded). Log(M*)9 9.5 10 10.50.20.40.60.811.2 Log(M*)9 9.5 10 10.50.20.40.60.811.2
Figure 12.
Fraction of galaxies with
Log ( SF R/M ∗ ) higher and D n (4000) lower than the median of the sample for central galaxies in sys-tems without interaction (red solid) and systems with interaction betweensatellites or with main galaxy (blue dotted) and control sample (shaded) formation activity. It can also be appreciated that this effect is morenoticeable for low stellar masses, and it decreases with negligibledifferences for the most massive objects.The contribution of the effect of each member of the systemcan be seen in the middle and right panels of Fig. 15, for the centralgalaxy and for the sum of its satellite stellar contributions respec-tively. It can be seen that, as in the previous section, the trendsare mainly dominated by the central galaxy. However, a significantincrease in the normalised SFR of low-mass satellites can be ob-served.Lambas et al. (2012) results for galaxy pairs in minor interac-tions are in general agreement with our findings. However, we no-tice a possible evolution scenario where lower mass systems, moresusceptible to environment, respond with significant bursts of starformation and associated colour changes. More massive systemswhich have more probably experienced previous interactions, aremore evolved thus resembling fossil groups. Using data from SDSS-DR14 we have built a sample of centralgalaxies accompanied by two satellites. We apply usually adoptedcriteria to define minor galaxy systems consistent with our previousstudies and in the literature. In addition we also impose an isolationcriterion to ensure that the identified systems are not affected bylarger structures.In order to study the presence and influence of interactionsin these small galaxy systems, we undertake a visual classificationprocedure where we considered cases with interactions betweensatellites, or between satellites and the central galaxy. Satellite andcentral samples were studied separately and for each of these sam-ples we constructed control samples from the galaxy catalogue hav-
MNRAS000
MNRAS000 , 1–14 (2020) atellites and central galaxies in SDSS Figure 13.
Examples of galaxy systems images with double interactions
20 40 60 80 100 120012345
20 40 60 80 100 120-10-10.5-11-11.5-12-12.5
Figure 14.
Left:Total star formation rate (
SF R c + SF R + SF R ) as a function of the radius of the minimal enclosing circle ( r mec ) for systems ofgalaxies without interaction (solid) and systems with interaction between satellites or with main galaxy (dotted). Dashed lines correspond to SF R c . Right: Log ( SF R total /M ∗ total ) vs ( r mec ) . Figure 15.
Top: Left: Total star formation rate (
SF R c + SF R + SF R ) as a function of total stellar mass ( M ∗ c + M ∗ + M ∗ ) for systems of galaxieswithout interaction (solid) and systems with interaction between satellites or with main galaxy (dotted). Centre: Star formation rate SF R c as a function ofstellar mass M ∗ c for central galaxies. Right: Total star formation rate ( SF R + SF R ) as a function of total stellar mass ( M ∗ + M ∗ ) for satellite galaxies.Bottom: same analysis for SFR normalised to M*.MNRAS , 1–14 (2020) Mesa et al. ing a similar z , M r and C distributions than those under consider-ation.These galaxy systems will be the basis of future observationalstudies. In the medium term we will conduct a study in multiplewavelengths, and expand the sample with other available cata-logues. In addition, these results may be used to make predictionsin future high redshift catalogues. The scientific aim of these newstudies is mainly focused at shedding light to the relevance of thedifferent mechanisms present in galaxies in close interactions.These process are key to understand the structural evolution aswell as changes in stellar populations and their impact on globalastrophysical characteristicsThe main results of our analysis can be summarised as fol-lows: • According to the selection criteria of the sample, two pop-ulations of very different galaxies have been obtained. The firstcomposed by the central galaxies, and the second by its satellites.Among them, there are notable differences in mass and brightness,as expected. • After the visual classification, it was found that around 80% ofthe systems do not show evident interactions, and that the remain-ing 20% does, and this mutual interaction may be between satellitesor any of them with their central galaxy. • The study of the central galaxies showed that this sampleis composed of more evolved galaxies, and therefore with red-der colours, old populations and with little star formation. This ismainly due to the nature of the sample chosen, given the criteriafor its selection, which puts certain restrictions on brightness. It isobserved that these luminous galaxies tend to be rather elliptical, orspirals with prominent bulges. • An analysis of the star formation and stellar populationsshowed that the systems that present interactions differ from therest, with signs of recent stellar formation and younger populations.Systems without interactions behave similarly to the control sam-ple. All these trends are decreasing as the galaxy’s mass increases,the greater the mass, the difference is not observed. • The analysis of the colours showed that in general and regard-less of the type of interaction, these galaxies tend to be rather red.With the control sample with some more dispersion. • The counterpart referred to satellite galaxies shows that thesegalaxies have a bimodal behaviour, with a part with old stellar pop-ulations and poor star formation, and on the other hand an impor-tant fraction of young and formative objects is observed. These val-ues correlate directly with the mass of each object, even the leastmassive ones that show these signs. This behaviour is independentof the type of interaction. The control sample is always showingyounger populations and greater stellar formation than the objectsthat are in the systems under study. • With regard to colours, it is observed that the satellite sampleis generally redder, and its control sample shows more blue colours. • We have considered particularly the case of double interac-tions with the central galaxy. The members of these systems showlarge star formation activity and young stellar populations and tracea tight colour-magnitude relation. However, it is necessary to in-crease the number of studied systems to confirm this trend. • In order to understand the observed trends globally, an analy-sis dependent on the projected distance r p was made, consideringa subsample with objects with r p less than the median value. Theresults found confirm the trends already observed, also highlightingthe incidence of r p in these. The results are very different for satel- lite and central galaxies. Highlighting mainly the central galaxiesbelonging to the subsample with shorter projected distances, wherethe fractions change significantly according to the interaction. Inthe total sample, the trend is maintained although to a lesser extent.On the opposite side are the satellites, in the total sample, wherethere are no notable differences in the percentages. If the subsam-ple of less than r p is taken into account, small differences can beseen. • We have studied the global star formation efficiency of thesystem and its dependence on total mass and on the radius of theminimal enclosing circle of the members ( r mec ) . We find a strongdependence of the total SFR on these parameters for systemsundergoing interactions.For all of the above, it can be concluded that both galaxy popu-lations studied in this work are already differentiated by nature, andthat makes their properties in general very different. Now, when in-teractions come into play, things begin to diversify, and show onceagain, that the effect affects in an unique way according to whatplace each one occupies in the system.While the central galaxies, by nature more red and passive(due to the selection constraints), when involved in an interactionrejuvenate and begin to show signs of recent star formation andyounger populations, those satellite galaxies do not show differ-ences in this aspect. This supports the idea that starbursts occur butmainly in the central galaxy. This is also evident when compared tothe control sample, galaxies with similar redshift, luminosity andmorphology, but isolated in this case, always show more stellar ac-tivity than their counterpart in our systems. La Barbera et al. (2014)showed that the star formation history of early-type central galaxieshave a significant dependence on the environment, being those be-longing to groups who show an stellar formation activity that lastsin time, driven by the constant encounters with their satellite galax-ies. Additionally, it has been observed that all these trends cor-relate directly with the mass and projected distance between themembers involved in the interaction. These results support previousfindings which shows that galaxy interactions are powerful mech-anisms to trigger starburst and modify different galaxy properties(e.g. Lambas et al. 2003; Alonso et al. 2012; Mesa et al. 2014;Knapen et al. 2015; Moreno et al. 2020). In addition, the massesand the closeness between galaxies involved in the merger are im-portant parameters in setting the effects of the interactions (Bartonet al. 2000; Ellison et al. 2008; Lambas et al. 2012). In this waymore closed and less massive systems show efficient starbursts re-flected in young stellar population and bluer colours. On the otherhand, more massive systems present truncated star formation ac-tivity indicating a more evolve state. This scenario suggests thatmassive systems may have experienced interactions in the past andcould be a previous stage of the fossil groups. ACKNOWLEDGEMENTS
We would like to thank to the anonymous referee for a detailed re-vision of the manuscript and for the suggestions that helped to im-prove this paper. This work was partially supported by the ConsejoNacional de Investigaciones Científicas y Técnicas and the Sec-retaría de Ciencia y Técnica de la Universidad Nacional de SanJuan. V.M. also acknowledges support from project Fondecyt No.3190736.Funding for the SDSS has been provided by the Alfred
MNRAS , 1–14 (2020) atellites and central galaxies in SDSS DATA AVAILABILITY
The data underlying this article will be shared on reasonable re-quest to the corresponding author.
REFERENCES
Abolfathi B., et al., 2018, ApJS, 235, 42Abraham R. G., Valdes F., Yee H. K. C., van den Bergh S., 1994, ApJ, 432,75Alonso M. S., Tissera P. B., Coldwell G., Lambas D. G., 2004, MNRAS,352, 1081Alonso M. S., Lambas D. G., Tissera P., Coldwell G., 2007, MNRAS, 375,1017Alonso S., Mesa V., Padilla N., Lambas D. G., 2012, A&A, 539, A46Balogh M. L., Morris S. L., Yee H. K. C., Carlberg R. G., Ellingson E.,1999, ApJ, 527, 54Balogh M., et al., 2004, MNRAS, 348, 1355Barber C., Starkenburg E., Navarro J. F., McConnachie A. W., 2015, MN-RAS, 447, 1112Barrow J. D., Bhavsar S. P., Sonoda D. H., 1984, MNRAS, 210, 19PBarton E. J., Geller M. J., Kenyon S. J., 2000, ApJ, 530, 660Bernardi M., 2009, MNRAS, 395, 1491Blanton M. R., Roweis S., 2007, AJ, 133, 734Blanton M. R., Eisenstein D., Hogg D. W., Schlegel D. J., Brinkmann J.,2005, ApJ, 629, 143Blanton M. R., et al., 2017, AJ, 154, 28Brinchmann J., Charlot S., White S. D. M., Tremonti C., Kauffmann G.,Heckman T., Brinkmann J., 2004, MNRAS, 351, 1151Daddi E., et al., 2005, ApJ, 626, 680De Lucia G., Hirschmann M., Fontanot F., 2019, MNRAS, 482, 5041Deason A., Wetzel A., Garrison-Kimmel S., 2014, ApJ, 794, 115Deng X.-F., 2013, Research in Astronomy and Astrophysics, 13, 651Duplancic F., Coldwell G. V., Alonso S., Lambas D. G., 2018, MNRAS,481, 2458Ellison S. L., Patton D. R., Simard L., McConnachie A. W., 2008, AJ, 135,1877Ellison S. L., Patton D. R., Simard L., McConnachie A. W., Baldry I. K.,Mendel J. T., 2010, MNRAS, 407, 1514Ellison S. L., Patton D. R., Mendel J. T., Scudder J. M., 2011, MNRAS,418, 2043 Forbes D. A., Romanowsky A. J., Pastorello N., Foster C., Brodie J. P.,Strader J., Usher C., Pota V., 2016, MNRAS, 457, 1242Gadotti D. A., 2009, MNRAS, 393, 1531Gutiérrez C. M., Alonso M. S., Funes J. G., Ribeiro M. B., 2006, AJ, 132,596Hirschmann M., De Lucia G., Wilman D., Weinmann S., Iovino A., CucciatiO., Zibetti S., Villalobos Á., 2014, MNRAS, 444, 2938Hirschmann M., Naab T., Ostriker J. P., Forbes D. A., Duc P.-A., Davé R.,Oser L., Karabal E., 2015, MNRAS, 449, 528Hogg D. W., et al., 2004, ApJL, 601, L29Kauffmann G., et al., 2003, MNRAS, 341, 33Kennicutt Robert C. J., 1998, ARA&A, 36, 189Knapen J. H., Cisternas M., Querejeta M., 2015, MNRAS, 454, 1742Knobel C., et al., 2013, ApJ, 769, 24La Barbera F., Pasquali A., Ferreras I., Gallazzi A., de Carvalho R. R., de laRosa I. G., 2014, MNRAS, 445, 1977Lambas D. G., Tissera P. B., Alonso M. S., Coldwell G., 2003, MNRAS,346, 1189Lambas D. G., Alonso S., Mesa V., O’Mill A. L., 2012, A&A, 539, A45Mesa V., Duplancic F., Alonso S., Coldwell G., Lambas D. G., 2014, MN-RAS, 438, 1784Mesa V., Duplancic F., Alonso S., Muñoz Jofré M. R., Coldwell G., LambasD. G., 2018, A&A, 619, A24Moreno J., et al., 2020, MNRAS,Morselli L., Popesso P., Erfanianfar G., Concas A., 2017, A&A, 597, A97Naab T., Johansson P. H., Ostriker J. P., 2009, ApJL, 699, L178Nipoti C., Giocoli C., Despali G., 2018, MNRAS, 476, 705O’Mill A. L., Duplancic F., García Lambas D., Valotto C., Sodré L., 2012,MNRAS, 421, 1897Oman K. A., Hudson M. J., 2016, MNRAS, 463, 3083Oser L., Ostriker J. P., Naab T., Johansson P. H., Burkert A., 2010, ApJ,725, 2312Pasquali A., Smith R., Gallazzi A., De Lucia G., Zibetti S., Hirschmann M.,Yi S. K., 2019, MNRAS, 484, 1702Patton D. R., Torrey P., Ellison S. L., Mendel J. T., Scudder J. M., 2013,MNRAS, 433, L59Patton D. R., Qamar F. D., Ellison S. L., Bluck A. F. L., Simard L., MendelJ. T., Moreno J., Torrey P., 2016, MNRAS, 461, 2589Patton D. R., et al., 2020, MNRAS, 494, 4969Perez J., Tissera P., Blaizot J., 2009, MNRAS, 397, 748Rogers B., Ferreras I., Kaviraj S., Pasquali A., Sarzi M., 2009, MNRAS,399, 2172Sales L. V., et al., 2015, MNRAS, 447, L6Shao S., Cautun M., Frenk C. S., Grand R. J. J., Gómez F. A., MarinacciF., Simpson C. M., 2018, MNRAS, 476, 1796Sol Alonso M., Lambas D. G., Tissera P., Coldwell G., 2006, MNRAS, 367,1029Stierwalt S., Besla G., Patton D., Johnson K., Kallivayalil N., Putman M.,Privon G., Ross G., 2015, ApJ, 805, 2Strateva I., et al., 2001, AJ, 122, 1861Tempel E., Guo Q., Kipper R., Libeskind N. I., 2015, MNRAS, 450, 2727Tinker J. L., Wetzel A. R., 2010, ApJ, 719, 88Trujillo I., et al., 2006, MNRAS, 373, L36Trujillo I., Conselice C. J., Bundy K., Cooper M. C., Eisenhardt P., EllisR. S., 2007, MNRAS, 382, 109Vulcani B., et al., 2014, MNRAS, 441, 1340Weinmann S. M., van den Bosch F. C., Yang X., Mo H. J., 2006, MNRAS,366, 2Wetzel A. R., Tinker J. L., Conroy C., 2012, MNRAS, 424, 232Wetzel A. R., Tinker J. L., Conroy C., van den Bosch F. C., 2013, MNRAS,432, 336Woods D. F., Geller M. J., 2007, AJ, 134, 527Yamauchi C., et al., 2005, AJ, 130, 1545Yee H. K. C., Ellingson E., 1995, ApJ, 445, 37York D. G., Adelman J., et al. A., 2000, AJ, 120, 1579van Dokkum P. G., et al., 2008, ApJL, 677, L5van den Bosch F. C., Aquino D., Yang X., Mo H. J., Pasquali A., McIntoshD. H., Weinmann S. M., Kang X., 2008, MNRAS, 387, 79MNRAS , 1–14 (2020) Mesa et al. van der Wel A., et al., 2014, ApJ, 788, 28This paper has been typeset from a TEX/L A TEX file prepared by the author. MNRAS000