Global environmental effects versus galaxy interactions

Abstract

We explore properties of close galaxy pairs and merging systems selected from the SDSS-DR4 in different environments with the aim to assess the relative importance of the role of interactions over global environmental processes. For this purpose, we perform a comparative study of galaxies with and without close companions as a function of local density and host-halo mass, carefully removing sources of possible biases. We find that at low and high local density environments, colours and morphologies of close galaxy pairs are very similar to those of isolated galaxies. At intermediate densities, we detect significant differences, indicating that close pairs could have experienced a more rapid transition onto the red sequence than isolated galaxies. The presence of a correlation between colours and morphologies indicates that the physical mechanism responsible for the colour transformation also operates changing galaxy morphologies. Regardless of dark matter halo mass, we show that the percentage of red galaxies in close pairs and in the control sample are comparable at low and high local density environments. However, at intermediate local densities, the gap in the red fraction between close pairs and the control galaxies increases from ~10% in low mass haloes up to ~50% in the most massive ones. Our findings suggest that in intermediate density environments galaxies are efficiently pre-processed by close encounters and mergers before entering higher local density regions. (Abridge)

Full text

Page 1

arXiv:0904.2851v2 [astro-ph.CO] 21 Apr 2009
Mon. Not. R. Astron. Soc. 000, 000–000 (0000)Printed 21 April 2009(MN LATEX style file v2.2)
Global environmental effects versus galaxy interactions
Josefa Perez1,2,3⋆, Patricia Tissera1,3, Nelson Padilla4, M. Sol Alonso3,5and Diego G. Lambas
1Instituto de Astronom´ ıa y F´ ısica del Espacio,Conicet-UBA, CC67, Suc.28,Ciudad de Buenos Aires, Argentina.
2Facultad de Ciencias Astronom´ ıa y Geof´ ısica, Universidad Nacional de La Plata, Argentina.
3Consejo Nacional de Investigaciones Cient´ ıficas y T´ ecnicas, CONICET, Argentina.
4Departamento de Astronom´ ıa y Astrof´ ısica, Pontificia Universidad Cat´ olica de Chile, Santiago, Chile.
5Complejo Astron´ omico El Leoncito, CP J5402DSP, San Juan, Argentina.
6Observatorio Astron´ omico de la Universidad Nacional de C´ ordoba, Argentina.
21 April 2009
ABSTRACT
We explore properties of close galaxy pairs and merging systems selected from the
SDSS-DR4 in different environments with the aim to assess the relative importance
of the role of interactions over global environmental processes. For this purpose, we
perform a comparative study of galaxies with and without close companions as a
function of local density and host-halo mass, carefully removing sources of possible
biases.
We find that at low and high local density environments, colours and morpholo-
gies of close galaxy pairs are very similar to those of isolated galaxies. At intermediate
densities, we detect significant differences, indicating that close pairs could have ex-
perienced a more rapid transition onto the red sequence than isolated galaxies. The
presence of a correlation between colours and morphologies indicates that the physi-
cal mechanism responsible for the colour transformation also operates changing galaxy
morphologies. At fixed local densities, we find a dependence of the red galaxy fraction
on dark matter halo mass for galaxies with or without a close companion. This sug-
gests the action of host halo mass related effects. Regardless of dark matter halo mass,
we show that the percentage of red galaxies in close pairs and in the control sample
are comparable at low and high local density environments. However, at intermediate
local densities, the gap in the red fraction between close pairs and the control galaxies
increases from ∼ 10% in low mass haloes up to ∼ 50% in the most massive ones.
Interestingly, we also detect that 50% of merging systems populate the intermediate
local environments, with a large fraction of them being extremely red and bulge dom-
inated. Our findings suggest that in intermediate density environments galaxies are
efficiently pre-processed by close encounters and mergers before entering higher local
density regions.
Key words: galaxies: evolution, galaxies: general, galaxies: interactions.
1 INTRODUCTION
Several observational and theoretical works have gathered
evidence to determine that the environment where galax-
ies reside plays a fundamental role in shaping their prop-
erties. However, although the transformation of blue, late-
type and star-forming field galaxies into red, early-type and
passive cluster galaxies have been well established (Oemler
1974; Dressler 1980; Lewis et al. 2002; Gomez et al. 2003;
Balogh et al. 2004; Baldry et al. 2004; O’Mill, Padilla &
Lambas, 2008), there is no consensus on the mechanisms re-
⋆E-mail: jperez@fcaglp.unlp.edu.ar
sponsible for this transformation. Many explanations have
been proposed including: i) ram-pressure stripping of cold
interstellar gas of galaxies falling at high velocities into the
ICM, which produces a fast truncation of the star formation
(Gunn & Gott 1972); ii) starvation or strangulation, which
are also stripping gas processes of the hot diffuse compo-
nent of satellite galaxies, which affects the star formation
on longer timescales (Larson et al. 1980); iii) harassment,
the cumulative effect of several rapid encounters with other
cluster members, which leads to substantial changes in the
galaxy morphology (Moore et al. 1998); iv) mergers and in-
teractions of galaxies which can trigger an intense burst of
star formation, rapidly consuming the cold gas and forming

Page 2

2
Perez et al.
spheroidal systems (Toomre & Toomre 1972, Kauffmann et
al. 1993). These explanations, however, are still under dis-
cussion.
van den Bosch et al. (2008) and Weinmann et al. (2008)
analyse the role of satellite quenching for the build-up the
red galaxy sequence. They find that the environmental pro-
cesses which shut down the star formation (SF) activity in
satellite galaxies are equally efficient in host haloes of all
masses. This rules against mechanisms that are thought to
operate only in very massive haloes, such as ram-pressure or
harassment. They suggest that the process responsible for
quenching the SF in satellites should last a timescale of a
few Gyr, suggesting starvation as the satellite-specific trans-
formation mechanism. They also claim that an additional
mechanism is also required because quenching alone cannot
explain the morphological transformations in the build-up
of the red sequence.
Alternatively, attempts have been made to remove the
problem from clusters entirely by proposing a preprocessing
of disc blue galaxies to red earlier type systems at moderate
environments (Balogh et al. 2004; Mihos 2004; Moss 2006;
Patel et al. 2008). Recently, many authors have investigated
the dependence of galaxy properties on environment at in-
termediate densities (i.e. galaxy groups in the outskirt of
clusters, infall populations) suggesting different mechanisms
to account for them. Patel et al. (2008) find that galaxies at
cluster-centric radii larger than 3 Mpc show an enhanced red
galaxy fraction, indicating that intermediate density regions
and groups in the outskirts of clusters are locations where
the local environment influences the transition of galaxies
onto the red-sequence, as opposed to mechanisms that op-
erate on cluster scales (e.g. ram-pressure stripping, harass-
ment). In the same direction, Moss (2006) provides evidence
stating that the cluster giant S0 population can be explained
as the outcome of minor mergers with the infalling popula-
tion integrated over the past ∼ 10 Gyr.
Analysis of the SF at intermediate environments show
that the current SFR of a galaxy falling along a supercluster
filament is likely to undergo a sudden enhancement before
the galaxy reaches the virial radius of the cluster (Porter
et al. 2008). These authors suggest that the main process
responsible for this rapid burst are close interactions with
other galaxies in the same filament, if the interactions occur
before the gas reservoir of the galaxy gets stripped off due
to the interaction with the ICM.
On the other hand, Gallazzi et al. (2008) explore the
amount of obscured SF as a function of environments in the
A901/902 supercluster. The SF hidden among red galaxies
is detected by using SF indicators that are not affected by
dust attenuation. Otherwise, they could be missed or mis-
takenly classified as post-starburst on the basis of their weak
emission lines obtained via optical, dust-sensitive SFR indi-
cators. Combining the near UV/optical data with infrared
photometry, they find that ∼ 60% of red star-forming galax-
ies have IR-to-UV luminosity ratios which indicate high dust
obscuration. Interestingly, most of them populate interme-
diate density regions. In agreement with this result, Wolf
et al. (2005) have also identified an excess of dusty red
galaxies with young stellar populations in the infalling re-
gion of the same cluster. More evidence of dust-obscured
SF at intermediate regions is provided by Miller & Owen
(2002). They find that up to ∼ 20% of the galaxies in 20
nearby Abell clusters have dust-obscured SF and are pref-
erentially located at intermediate density regions, with re-
spect to normal star-forming galaxies or AGNs. In addi-
tion, Poggianti et al. (2008) report that dusty starburst can-
didates present a very different environmental dependence
than post-starburst galaxies. They find that the spectra of
dusty starburst candidates are numerous in all environments
at intermediate redshifts, particularly among galaxy groups.
This favours the hypothesis of dusty starbursts triggered by
mergers, expected to be common in groups.
Motivated by these findings, in this paper, we revise the
role of mergers and galaxy interactions in driving galaxy evo-
lution at different density environments by using the SDSS-
DR4 data. In a hierarchical clustering universe, as Mihos
(2004) noticed, galaxy clusters would form not by accreting
individual galaxies from the field, but rather through the in-
fall of less massive groups moving along the filaments. Such
infalling groups provide locations with much lower velocity
dispersions than the cluster medium, thus permitting strong
slow encounters more normally associated with the field. In
agreement, Moss (2006) shows that ∼ 50% − 70% of the in-
fall population are found to be in merging systems and slow
galaxy-galaxy encounters. Hence, we will focus our analysis
on close encounters and merger candidates in order to dis-
entangle their role at such intermediate density regions. But
before going any further, we analyse a technical issue.
The effect of galaxy interactions has been largely stud-
ied in optical and infrared surveys (Lambas et al. 2003;
Nikolic et al. 2004; Alonso et al. 2004; De Propris et al.
2005; Kewley et al. 2006; Alonso et al. 2006; Lin et al. 2007;
Michel-Dansac et al. 2008; Park et al. 2009) and also us-
ing numerical simulations (Barnes & Hernquist 1996; Mi-
hos & Hernquist 1996; Larson & Tinsley 1978; Perez et al.
2006a,b). These results conclude that mergers and close in-
teractions could actually modify the star formation, mor-
phology, colours and metallicities of galaxy pairs. In all
these cases, people have attempted to isolate the effects of
galaxy interactions by comparing galaxies in pairs with iso-
lated galaxies. However, different authors have proposed dif-
ferent ways to build these control samples (CS). By using
SDSS-DR4 mock galaxy catalogues built using the Millen-
nium Simulation, Perez et al. (2009, hereafter P09) show
that the set of constrains used to define a CS might intro-
duce biases which could affect the interpretation of results.
The fact that the physics of interactions is not included in
the semi-analytic model (SAM) studied in P09 allows the au-
thors to attribute differences between the mock control and
pair samples solely to selection biases. P09 find that these
biases are diminished by ≈ 70% after imposing constrains
on redshifts, stellar masses and local densities, and are com-
pletely removed when halo masses are also considered (also
see Barton et al. 2007). However, P09 notice that the con-
tribution of the halo mass bias introduced by the recipes
adopted in SAMs to model satellite galaxies is probably ex-
acerbated. Based on these previous theoretical findings, we
first build an unbiased CS from the SDSS-DR4 data in or-
der to isolate signals of galaxy interactions in a more robust
way.
This paper is organized as follows. In Section 2, we de-
scribe the SDSS-DR4 Galaxy Pair Catalogue and CS. We
analyse possible bias effects in the selection of CS (2.1) and
correct them in order to obtain an unbiased CS, suitable for

Page 3

3
isolating the effect of galaxy interactions (2.2). In Section
2.3, we revise some of the previous observational results of
galaxy pairs by comparing them with isolated galaxies us-
ing the unbiased CS. The role of galaxy interactions and
mergers in different local density and host-halo mass envi-
ronments are discussed in Section 3. Finally, in Section 4 we
summarize our results.
2 BUILDING A CONTROL SAMPLE FOR A
GALAXY PAIR CATALOGUE.
The analysis of this paper is based on the SDSS-DR4 photo-
metric and spectroscopic galaxy catalogue, for which there
are estimations of gas-phase oxygen abundances, stellar
masses and metallicities provided by Tremonti et al. (2004)
and Gallazzi et al. (2005). Star formation rate estimates
for galaxies are obtained as described in Brinchmann et al.
(2004). The SDSS-DR4 galaxy sample is essentially a mag-
nitude limited spectroscopic sample with rlim < 17.77 cov-
ering a redshift range 0 < z < 0.25. We considered a shorter
redshift range, 0.01 < z < 0.1, in order to avoid strong in-
completeness at larger distances (Alonso et al. 2006). We
also excluded from our sample AGNs, which could affect
our interpretation of results due to contributions from their
emission line spectral features.
We characterized the local environment of galaxies
defining a projected density parameter, Σ. This parameter
is calculated by using the projected distance to the 5thnear-
est neighbour, Σ = 5/(πd2
neighbours have been chosen to have luminosities brighter
than Mr < −20.5 and radial velocity differences lower than
1000kms−1. In the case of pairs, we also estimated the local
density by using the 6thnearest neighbour, in order to as-
sess possible biases in the definition of the local density for
galaxy pairs. However, we find that our results are insensi-
tive to the definition of Σ by using either the 5thor the 6th
nearest neighbour.
We use the SDSS-DR4 galaxy group catalogue from Za-
pata et al. (2009) to assign a host halo mass (Mvir) to each
individual galaxy in our sample. This group catalogue is
complete above masses of 1013h−1M⊙, out to the limit red-
shift of our galaxy sample, z = 0.1. The mass assignment is
done by searching for the closest group, in terms of its virial
radius rvir, to each galaxy; if a group is found within 1.5 rvir
in projection and with a velocity difference, ∆v < 1000km/s
the galaxy is assigned the group mass as its host halo mass.
Galaxies which do not satisfy these conditions for any group
are assumed to be hosted by haloes below the mass detection
limit of 1013h−1M⊙.
Following Alonso et al. (2006, and references therein),
we build a Galaxy Pair Catalogue (GPC) requiring members
to have relative projected separations rp < 100kpch−1and
relative radial velocities ∆V < 350kms−1. Galaxies without
a near companion within the adopted thresholds constitute
the Isolated Galaxy Sample (IGS). In order to properly as-
sess the significance of the results obtained from the GPC,
we use a control sample (CS) constructed by selecting galax-
ies from the IGS. In a first attempt to build a CS, we follow
recent observational works (Lambas et al., 2003; Alonso et
al., 2006; Michel-Dansac et al., 2008) and define a CS by
selecting galaxies from the IGS which match one-to-one the
5). Following Balogh et al. (2004),
redshift and r-band magnitude of each galaxy in pairs, here-
after the SDSS-CS1.
However, the analysis of SDSS-DR4 mock catalogues
built using the Millennium Simulation shows that a CS de-
fined by only applying redshift and luminosity requirements
exhibits different stellar masses, morphologies, halo masses
and local projected density distributions than those of galax-
ies in pairs (P09) as discussed in the Introduction. Hence,
in this section, and based on these theoretical findings, we
analyse how these biases could affect the selection of the
SDSS-CS1, and then we correct them in order to build up
a robust ’unbiased’ control sample.
2.1Analysis of biases in the selection of the
SDSS-CS1.
In this subsection, we apply the tests devised by P09 to the
SDSS-CS1, excluding the morphological selection, which we
will not consider in our work since in the case of merging
systems or close galaxy pairs with tidal features, it is rather
difficult to objectively define their morphology.
In Fig. 1, we perform a comparative analysis of stel-
lar masses, halo masses and local projected density (Σ) dis-
tributions for SDSS-DR4 galaxy pairs (solid lines) and for
the SDSS-CS1 (dashed lines). The most significant differ-
ences between control and pair properties are observed for
the distributions of local density environment. In agreement
with theoretical findings by P09, galaxy pairs tend to inhabit
higher density regions than galaxies in the SDSS-CS1.
Consistently with P09, we find that galaxy pairs ex-
hibit slightly larger stellar masses than their isolated coun-
terparts in the SDSS-CS1. However, observations show that
pair and control halo mass distributions are similar, in con-
tradiction to P09 who reported that in the SAM the halo
mass is the main factor contributing to bias the mock-CS1.
As a matter of fact, P09 found that mock galaxy pairs are
hosted by larger dark matter haloes than galaxies in the
mock-CS1. However, they warned that this effect could be
overestimated in the SAM due to the environmental treat-
ment of the starvation of hot gas in satellite galaxies (Wein-
mann et al. 2006; Kang et al. 2008; P09). Thus, the fact that
SDSS galaxies with and without a near companion have sim-
ilar halo mass distributions (Fig. 1), could be considered as
an indication of the exacerbated environmental modelling of
satellite galaxies in the SAM. However, we remind the reader
that the dynamical estimates of halo masses used in this
present paper could also be introducing potential sources of
errors due to the uncertainties introduced by velocity dis-
persions (Eke et al. 2004; Padilla et al. 2004).
The observational results shown in Fig. 1 indicate that
even when the local density environment is the most signifi-
cant bias in the SDSS-CS1, we should also take into account
both stellar and halo masses in order to properly select the
SDSS control sample.
2.2Treatment of selection effects
In order to remove the bias introduced by the differences in
stellar masses, we define a CS using stellar masses instead
of magnitudes as a constrain. Thus, the SDSS-CS2 is built
up by selecting galaxies from the IGS so that the distribu-
tions of redshifts and stellar masses match those of galaxies

Page 4

4
Perez et al.
Figure 1. Distributions of local density parameter (Σ) (upper
panel), stellar masses (middle panel) and dark matter halo masses
(lower panel), for SDSS galaxies in pairs (solid lines) and in the
SDSS-CS1 (dashed lines).
in pairs. Ellison et al. (2008) also build a CS using a similar
approach to our SDSS-CS2 in order to study close galaxy
pairs in the SDSS. Additionally, we build a SDSS-CS3 im-
posing constrains on redshift, stellar mass and halo mass.
Finally, we define the SDSS-CS4 populated by IGS galaxies
matching the galaxy pair redshift, stellar mass, local density
environment and halo masses; the SDSS-CS4 should not be
affected by selection biases according to the results by P09.
After this CS building process, we have to remark that
in order to construct the more suitable and ’unbiased’ SDSS-
CS4, we remove around 8% of the galaxies in pairs from the
original GP sample to account for the lack of counterparts
in the IGS with identical stellar masses, local density envi-
ronments and halo masses. Most of the removed galaxies in
pairs reside in high density regions where galaxies satisfying
our isolation requirements are less common (Fig. 1).
2.3 Isolating the effect of interactions.
We have shown that the set of constrains used to define the
SDSS-CS1 introduces biases which could affect the interpre-
tation of the results found for galaxy pairs. In this section,
we revise some previous observational analysis of galaxy
pairs from the SDSS-CS1 (Lambas et al. 2003; Alonso et
al. 2006; Michel-Dansac et al. 2008) to evaluate how their
results would be modified considering the different CSs de-
fined in the previous subsection.
In order to test the performance of each CS, we study
their colours, metallicities and star formation activity, com-
paring them to those of galaxies in pairs. Particularly, we
analyse the dependence of the star formation activity on
the environment, as well as u − r colour distributions for
the SDSS-CS1, SDSS-CS2 and SDSS-CS3, finding simi-
lar results for all these CS (see P09 for a discussion on the
mass-metallicity relation). The most significant change is
detected when analysing the properties of SDSS-CS4, in-
dicating that, the local density environment is responsible
for introducing an important bias effect (see upper panel
of Fig.1). Using results from a semi-analytical model, P09
showed that the halo mass is the parameter with the largest
contribution towards biasing a CS. However, they could not
separate this latter bias effect from the modelling bias asso-
ciated to the treatment of satellites. The analysis of SDSS
control samples shows a less significant role of the dark mat-
ter mass in the CS definition than previously reported (P09,
Barton et al. 2007), and contributes to show the exacerbated
environmental modelling of satellite galaxies in the SAM
(Weinmann et al. 2006; Kang et al. 2008).
2.3.1 Comparative analysis of galaxy pairs, SDSS-CS2
and SDSS-CS4
Since bias effects are removed only after defining the SDSS-
CS4, from this point on we will only show the results for
samples SDSS-CS2 and the ’unbiased’ case. Note that we
have chosen the SDSS-CS2 instead of the SDSS-CS1, since
they behave similarly and, on the other hand, the stellar
mass is a more fundamental quantity than the luminosity of
a galaxy (Brinchmann et al. 2000; Kauffmann et al. 2003;
Panter et al. 2004).
We first analyse the SF activity in different den-
sity environments. Particularly, we estimate the SF his-
tory of systems by defining the stellar birthrate parameter,
b = 0.5tH(SFR/M∗), where tH is the Hubble time, and
SFR/M∗ is the present star formation rate normalized to
the total stellar mass (Brinchmann et al. 2004). In Fig. 2, we
show the contour plots of the stellar birthrate parameter as
a function of the local density estimator (Σ) for galaxy pairs
(left panels) and CSs (right panels). The upper plots show
the result for galaxies with and without a near companion
using the ’biased’ CS21definition. Lower panels show the
’unbiased’ results obtained by comparing galaxies in pairs
with those in the CS4. Note, that in order to define the
’unbiased’ CS4, we remove a small fraction of galaxies in
pairs, thus the galaxy pair samples also change from the
1For simplicity, hereafter from this point on we have suppressed
the SDSS prefix in sample names.

Page 5

5
top to the bottom panels. In order to clarify the discussion,
from this point on we will refer to the galaxy pair samples as
GP2 (upper left panel) and GP4 (lower left panel), respec-
tively. The green lines in the plots divide the three different
environmental regions discussed throughout this paper (see
Table 1 for description).
As we can appreciate in the top panels of Fig. 2, the
biased samples show that while galaxies with a near com-
panion (top left) have a clearer bimodal distribution in the
SF-environment relation (i.e. a large fraction of passive SF
galaxies populating high density regions), isolated galaxies
(top right) are more consistent with a unimodal SF distribu-
tion shifted to high values of the stellar birthrate parameter.
These significant differences between galaxies in GP2 and
in CS2, however, cannot be only attributed to galaxy in-
teractions, but also to a biased selection of these samples.
Indeed, we can see that the SF-environment relation of the
’unbiased’ CS4 (bottom right) significantly changes with
respect to that of CS2 (top right), resembling more closely
that of its pair counterpart, GP4 (bottom left). This means
that after correcting for the biases, the differences between
the SF-environment distributions of galaxies with and with-
out a near companion are partially reduced. However, dis-
crepancies still remain even when GP4 is compared to the
unbiased CS4, suggesting a real effect coming from galaxy
interactions. We find that, at high densities, the GP4 has
a larger fraction of low star-forming systems than the CS4
while at intermediate and low density environments, it has
larger fractions of strong star-forming galaxies.
Alonso et al. (2006) have previously analysed the effi-
ciency of galaxy interactions in driving the SF, and how the
environment can modify this efficiency. They show that, at
low and intermediate density regions, close galaxy interac-
tions are more effective at triggering important SF activity
than galaxies without a near companion. We revised their
estimates using samples 2 (CS2 and GP2) and samples 4
(CS4 and GP4) for the purpose of assessing possible bias
effects. Fig. 3 shows the fraction of strong SF galaxy pairs
as a function of their relative projected separations, rp. This
fraction is defined by systems with SF activity larger than
the mean b of their corresponding control sample so that
they have b > 1. The analysis was performed at high (dashed
lines), intermediate (solid lines) and low (dotted lines) den-
sity environments (see Table 1). Horizontal lines represent
the mean value of the fraction associated to the correspond-
ing CS in each environment. We reproduce the findings of
Alonso et al. (2006) where the only significant change is de-
tected in low density regions, where we find that galaxy pairs
are required to be closer than in GP2 (within 1σ) in order
to exhibit an enhanced SF activity. This is a consequence of
an increment of the SF activity in the CS4 with respect to
that measured in the CS2 at low densities, indicated by the
increase of the fraction of b > 1 galaxies from 0.25 to 0.29
(see also Fig. 2 (right panels)).
Given that galaxy pairs closer than a critical relative
projected separation are more efficient at forming stars than
their isolated counterparts, we will concentrate our analysis
on studying the properties of close galaxy pairs from this
point on. The lower panel of Fig. 3 shows that regardless
of environment, systems with rp < 20kpch−1have a sta-
tistically significant enhancement in their SF activity. Thus,
Figure 3. Fraction of strong star forming galaxies (b > 1 where
b has been normalized to the mean b of the corresponding control
sample) in the GP catalogue as a function of their relative pro-
jected separation, rp, at high (dashed lines), intermediate (solid
lines) and low (dotted lines) density environments (see Table 1).
Horizontal lines represent the fractions associated to each CS esti-
mated with respect to their mean SF activity. In the upper panel
we show the result for GP2 and CS2, and in the lower one, for
GP4 and CS4.
we select close galaxy pairs by using this threshold on the
relative projected separation.
Regarding colour distributions, observational results
(e.g. Alonso et al. 2006; De Propris et al. 2005) have re-
ported an excess of blue galaxies in close pairs with respect
to that found in their CS, associated with a larger frac-
tion of actively star-forming galaxies. They also find a larger
fraction of red galaxies in close pairs with respect to those
systems without a near companion, an effect that might in-
dicate that dust, stirred up during encounters, could affect
colours, partially obscuring the tidally-induced SF (Gallazzi
et al. 2008). Other possible interpretation of this trend is
that many galaxies in pairs have been very efficient at form-
ing stars during the early stages of their evolution, so that
at present they exhibit red colours. In Fig. 4, we can see that
even though the differences at the red peak are reduced af-
ter removing the bias effects in the selection of the CS4, the
excess of close galaxy pairs in both red and blue tails with
respect to the CS4 still persists, supporting the claim that
these trends are actually produced by galaxy interactions
and not introduced by a biased selection.