Spirals, Bridges, and Tails: A GALEX UV Atlas of Interacting Galaxies

Abstract

We have used the GALEX ultraviolet telescope to study stellar populations and star formation morphology in a well-defined sample of 42 nearby optically-selected pre-merger interacting galaxy pairs. Galaxy interactions were likely far more common in the early Universe than in the present, thus our study provides a nearby well-resolved comparison sample for high redshift studies. We have combined the GALEX NUV and FUV images with broadband optical maps from the Sloan Digitized Sky Survey to investigate the ages and extinctions of the tidal features and the disks. The distributions of the UV/optical colors of the tidal features and the main disks of the galaxies are similar, however, the tidal features are bluer on average in NUV - g when compared with their own parent disks, thus tails and bridges are often more prominent relative to the disks in UV images compared to optical maps. This effect is likely due to enhanced star formation in the tidal features compared to the disks rather than reduced extinction, however, lower metallicities may also play a role. We have identified a few new candidate tidal dwarf galaxies in this sample. Other interesting morphologies such as accretion tails and `beads on a string' are also seen in these images. We also identify a possible `Taffy' galaxy in our sample, which may have been produced by a head-on collision between two galaxies. In only a few cases are strong tidal features seen in HI maps but not in GALEX. Comment: Accepted by the Astronomical Journal. Figures 1-18 in color jpg at http://www.etsu.edu/physics/bsmith/research/sg/galex_SDSS/galex_SDSS.html

Full text

Page 1

arXiv:1001.0989v1 [astro-ph.CO] 6 Jan 2010
Accepted by the Astronomical Journal.
Preprint typeset using LATEX style emulateapj v. 11/10/09
SPIRALS, BRIDGES, AND TAILS: A GALEX UV ATLAS OF INTERACTING GALAXIES
Beverly J. Smith
Department of Physics and Astronomy, East Tennessee State University, Johnson City TN 37614
Mark L. Giroux
Department of Physics and Astronomy, East Tennessee State University, Johnson City TN 37614
Curtis Struck
Department of Physics and Astronomy, Iowa State University, Ames IA 50011
Mark Hancock
Department of Physics, University of California Riverside, Riverside CA 92521
Sabrina Hurlock
Department of Physics and Astronomy, East Tennessee State University, Johnson City TN 37614
Accepted by the Astronomical Journal.
ABSTRACT
We have used the GALEX ultraviolet telescope to study stellar populations and star formation
morphology in a well-defined sample of 42 nearby optically-selected pre-merger interacting galaxy
pairs. Galaxy interactions were likely far more common in the early Universe than in the present,
thus our study provides a nearby well-resolved comparison sample for high redshift studies. We have
combined the GALEX NUV and FUV images with broadband optical maps from the Sloan Digitized
Sky Survey to investigate the ages and extinctions of the tidal features and the disks. The distributions
of the UV/optical colors of the tidal features and the main disks of the galaxies are similar, however,
the tidal features are bluer on average in NUV − g when compared with their own parent disks, thus
tails and bridges are often more prominent relative to the disks in UV images compared to optical
maps. This effect is likely due to enhanced star formation in the tidal features compared to the disks
rather than reduced extinction, however, lower metallicities may also play a role. We have identified
a few new candidate tidal dwarf galaxies in this sample. Other interesting morphologies such as
accretion tails and ‘beads on a string’ are also seen in these images. We also identify a possible ‘Taffy’
galaxy in our sample, which may have been produced by a head-on collision between two galaxies. In
only a few cases are strong tidal features seen in HI maps but not in GALEX.
Subject headings: galaxies: starbursts — galaxies: interactions— galaxies: ultraviolet
1. INTRODUCTION
Tidal disturbances have played an important role in
reshaping galaxies and triggering star formation over
cosmic time (see Struck 1999 for review).
infrared, and mid-infrared observations show that the
mass-normalized star formation rates of pre-merger in-
teracting systems are enhanced by a factor of two on av-
erage compared to normal spirals (e.g., Bushouse 1987;
Kennicutt et al. 1987; Bushouse, Lamb, & Werner 1988;
Barton et al. 2000; Barton Gillespie, Geller, & Kenyon
2003; Smith et al. 2007; Lin et al. 2007). Further, closer
pairs have more enhanced star formation than wider
pairs, but with significant scatter (Barton et al. 2000;
Lambas et al. 2003; Nikolic, Cullen, & Alexander 2004;
Lin et al. 2007).This effect is more difficult to see
in broadband optical colors; Larson & Tinsley (1978)
Hα, far-
smithbj@etsu.edu
girouxm@etsu.edu
curt@iastate.edu
mhancock@ucr.edu
zshh7@goldmail.etsu.edu
found a larger scatter in the broadband optical col-
ors of Arp Atlas galaxies than isolated galaxies, while
Bergvall, Laurikainen, & Aalto (2003) found little color
difference between an interacting and a more isolated
sample.
Within interacting galaxies, star formation is often
enhanced in the nuclear regions compared to the disks
(Hummel et al.1990;Nikolic, Cullen, & Alexander
2004).Luminous star forming regions are some-
timesseenin tidalfeatures
Mirabel et al.1991,1992;
1996).For a sample of 25 Arp Atlas galaxies,
Schombert, Wallin, & Struck-Marcell (1990) found that
the B − V colors of the tidal features were somewhat
bluer on average than the parent disks, but with signif-
icant scatter. They conclude that this difference is due
to either enhanced star formation or lower extinction in
tidal features.
With the advent of the Galaxy Evolution Explorer
(GALEX), a new window on star formation in galaxies
is now available. The addition of UV helps to break the
(Schweizer1978;
Hibbard & van Gorkom

Page 2

2Smith et al.
age−extinction degeneracy in population synthesis mod-
eling (e.g., Smith et al. 2008). Furthermore, since the
UV traces somewhat older and lower mass stars (≤400
Myrs; O to early-B stars) than Hα (≤10 Myrs; early-
to mid-O stars), it provides a measure of star formation
over a longer timescale than Hα studies. GALEX imag-
ing of interacting systems have shown that tidal features
are sometimes quite bright in the UV (e.g., Neff et al.
2005). In some cases, tidal features previously thought
to be purely gaseous have been detected by GALEX (e.g.,
Hancock et al. 2007). In other systems, GALEX im-
ages have been used to identify new tidal features (e.g.,
Boselli et al. 2005).
To address these issues, we have used the GALEX
telescope to image a well-defined sample of more than
three dozen strongly interacting galaxies in the ultravi-
olet. Combined with broadband optical data, these im-
ages provide information about the star formation his-
tory and dust extinction within the galaxies. For four of
the galaxies in this sample (Arp 82, Arp 284, Arp 285,
and Arp 305), we have already published the GALEX
images as part of detailed studies of the distribution of
star formation within the galaxies, and compared with
numerical simulations of the interaction (Hancock et al.
2007, 2009; Smith et al. 2008; Peterson et al. 2009). The
GALEX images of Arp 24, 85, and 244 were previously
analyzed by Cao & Wu (2007), Calzetti et al. (2005),
and Hibbard et al. (2005), respectively. In the current
paper, we present the full GALEX dataset for the entire
sample, compare with optical data, discuss global and
tidal properties, and provide a brief discussion of each
system. In a followup paper, we compare with a sample
of normal galaxies.
The galaxies in our interacting sample were selected to
be relatively isolated binary systems, thus they are less
complex than many of the interacting systems studied re-
cently by GALEX, for example, the Hickson Group stud-
iesbyde Mello, Torres-Flores, & Mendes de Oliveira
(2008) and Torres-Flores et al. (2009), and the ram-
pressure-stripped NGC 5291 system (Boquien et al.
2007). Our galaxies were selected to be relatively simple
systems, thus are more amendable to numerical mod-
eling and detailed matching of simulations with multi-
wavelength datasets. In the current paper, we provide a
summary of the GALEX data for the full sample.
2. THE INTERACTING GALAXY SAMPLE
Our galaxy sample was selected from the Arp Atlas
of Peculiar Galaxies (Arp 1966), based on the follow-
ing criteria: 1) They are relatively isolated binary sys-
tems; we eliminated merger remnants, close triples, and
multiple systems in which the galaxies have similar opti-
cal brightnesses (systems with additional smaller angular
size companions were not excluded). 2) They are tidally
disturbed. 3) They have radial velocities less than <
10,350 km/s (most are < 3500 km/s). 4) Their total
angular size is>
∼3′, to allow for good resolution with
GALEX. One of our systems, Arp 297, consists of two
pairs at different redshifts. These two pairs are included
separately in our sample. We also include the nearby
pair NGC 4567, which fits the above criteria but is not
in the Arp Atlas.
After removing objects with too-bright stars in their
field that could not be avoided by shifting the target po-
sition, our sample consists of 42 pairs of galaxies. These
systems are listed in Table 1. Of these pairs, 13 were al-
ready reserved by GALEX guaranteed time projects. We
observed the remaining 29 systems (see Section 3). We
then combined our new observations with the archival
data for the other 13 systems that had already been ob-
served.
This GALEX sample overlaps with the sample that
we studied in the infrared using the Spitzer telescope
(Smith et al. 2007), however, it is not identical. Spitzer
observations were made of some of the galaxies that we
were not able to observe with GALEX because of UV-
bright stars in the field. In addition, the Spitzer survey
omitted galaxies with angular sizes of the individual disks
less than 30′′, while these galaxies were included in the
GALEX study.
3. OBSERVATIONS
Table 2 provides the dates, exposure times, and tile
names for the GALEX observations of our sample, in-
cluding both our new observations and the archival ob-
servations. When possible, we imaged each galaxy for
≥1500 seconds in both the far-ultraviolet (FUV) and
the near-ultraviolet (NUV) broadband filters of GALEX,
which have effective bandpasses of 1350 − 1705˚ A and
1750 − 2800˚ A, respectively. As shown in Table 2, some
systems were observed only in NUV or only in FUV, due
to a bright star in the field. For the archival study, we
only selected galaxies with at least 800 seconds total ex-
posure in either the FUV or NUV. The GALEX field of
view is circular with a diameter of 1.2 degrees. The pixel
size is 1.′′5, and the spatial resolution is ∼5′′. In some
cases, the GALEX target position was offset from the
position of the galaxy in order to avoid a nearby bright
star.
4. OTHER DATA
Of our 42 systems with GALEX data, 29 also have
broadband optical images available from the Sloan Dig-
itized Sky Survey (SDSS; Abazajian et al. 2003). These
galaxies are identified in Table 2. The SDSS ugriz fil-
ters have effective wavelengths of 3560˚ A, 4680˚ A, 6180˚ A,
7500˚ A, and 8870˚ A, respectively. Of our GALEX sam-
ple of 42 pairs, 31 have broadband Spitzer 3.6, 4.5, 5.8,
8.0, and 24 µm images available. Most of these have
been published in Smith et al. (2007). Of our 42 sys-
tems, 21 have published 21 cm HI maps, while 23 have
Hα maps available, either from the literature or our own
unpublished observations with the Southeastern Associ-
ation for Research in Astronomy (SARA) telescope. We
also have acquired new optical and Spitzer spectra of a
few of the star forming regions in some of the tidal fea-
tures in our sample. These will be discussed in later
papers (Hancock et al. 2010; Higdon et al. 2010).
5. OVERVIEW OF MORPHOLOGIES
In Figures 1 − 14 we show mosaics of the UV and
optical images for the sample galaxies that have both
GALEX and SDSS images. These galaxies are displayed
in order by Arp number, and are discussed individually
in Section 7. Figures 15 − 18 show the GALEX images
for the systems not observed by SDSS. Figures 1 − 18
show that the sample galaxies have a large range of col-
lisional morphologies, including M51-like systems, wide

Page 3

GALEX UV Atlas of Interacting Galaxies3
pairs with long tails and/or bridges, wide pairs with short
tails, close pairs with long tails, and close pairs with short
tails. Our sample also includes a possible ‘Taffy’ galaxy,
Arp 261, apparently produced by a near head-on colli-
sion between two equal-mass gas-rich galaxies, as well as
possible previously unidentified ring galaxies in Arp 112,
192, and 282.
Within the tidal features, we see a range of star for-
mation morphologies. In many systems, we see examples
of the so-called ‘beads on a string’ morphology, in which
regularly-spaced clumps of star formation are seen along
spiral arms and tidal features. These clumps are gener-
ally spaced about 1 kpc apart, which is the characteristic
scale for the gravitational collapse of molecular clouds
(Elmegreen & Efremov 1996). Examples of such ‘beads’
are seen in Arp 34, 35, 65, 72, 82, 84, 86, 100, 242, and
285.
In a few systems, we see very luminous star forming re-
gions at the base of a tidal feature. We call these features
‘hinge clumps’ (Hancock et al. 2009). These lie near the
intersection of the spiral density wave in the inner disk
and the material wave in the tail. These may form when
dense material in the inner disk gets pulled out into a
tail. This lowers the shear, which may allow more mas-
sive clouds to gravitationally collapse. Hinge clumps are
seen in Arp 65, 72, 82, 242, 270, and 305.
Our sample also includes some candidate ‘tidal dwarf
galaxies’ (TDGs), massive concentrations of young stars
near the tips of tidal features. The prototypical TDG in
the northern tail of Arp 105 (Duc et al. 1997) is included
in our sample, along with the well-studied TDGs in Arp
244 and Arp 245 (Mirabel et al. 1992; Duc et al. 2000).
In Arp 242, candidate TDGs are seen in both tails. In
Arp 112, in addition to the two main galaxies a third
fainter galaxy is seen, which may be either a TDG, a
background galaxy, a portion of a collisional ring, or a
pre-existing dwarf. In addition, we have identified pos-
sible TDGs in Arp 305 (Hancock et al. 2009), Arp 181,
and Arp 202. Faint UV clumps are also visible near the
end of the long HI tail south of Arp 270, but no optical
redshifts are available at present to confirm that these
are associated with the tail. These candidate TDGs are
discussed in detail in Section 7.
Our sample also contains numerous examples of ac-
cretion from one galaxy to another. One of the best-
studied examples is the northern tail of Arp 285, which
was likely produced from material accreted from the
southern galaxy (Toomre & Toomre 1972; Smith et al.
2008). According to our numerical simulations, the ma-
terial in this tail fell into the gravitational potential of
the northern galaxy, overshot that potential, and is now
gravitationally collapsing and forming stars (Smith et al.
2008). We call such features ‘accretion tails’, to distin-
guish them from classical tidal features. The inner west-
ern tail of Arp 284 was likely produced by the same pro-
cess (Struck & Smith 2003). The southern tail of Arp
105 may have formed by the same mechanism, as well
as the northwestern tail of Arp 34. Arp 269 may be an
example of accretion from one galaxy to another; alterna-
tively, it may be an example of ram pressure occurring
during the passage of one galaxy through the gaseous
disk of another. Arp 87 contains a good example of a
polar ring-like system, caused by accretion from a com-
panion. These systems are described in more detail in
Section 7.
Our sample contains only a few tidal features that have
high HI column densities (∼4 × 1020cm−2) but are not
detected in our GALEX maps or published optical maps:
Arp 84, Arp 269, Arp 270, and Arp 280. A few ad-
ditional tidal features have somewhat lower HI column
densities, between 6 × 1019cm−2and 1020cm−2, but
no GALEX/SDSS counterparts: Arp 85, 86, and 271.
Deeper GALEX and optical images are needed to check
whether these are truly starless structures. These fea-
tures are discussed in more detail in Section 7.
6. UV − OPTICAL COLORS
6.1. Photometry
Magnitudes in the various GALEX and SDSS bands for
the main disks and the tidal features are given in Table
3. For the systems with possible TDGs near the tips
of tidal features, we determined the colors of the TDG
separately from the connecting tail. These are labeled
‘TDG’ in Table 3, although we emphasize that in most
cases it is unclear whether these are truly TDGs. These
features are discussed in detail in Section 7.
To determine these magnitudes, we used a set of rect-
angular boxes that covered the observed extent of the tar-
geted area in the GALEX and SDSS images, but avoided
very bright stars. These boxes generally coincide with
those used in our Spitzer study (Smith et al. 2007), but
for some galaxies these were modified, for example, if
the observed extent of the tidal features or the disk was
larger in the GALEX images. For each system, the sky
was measured in a set of rectangular areas without bright
stars or galaxies. The uncertainties in Table 3 include
both the statistical uncertainty and an uncertainty due
to variations from sky region to region, as in Smith et al.
(2007). The magnitudes in Table 3 were corrected for
Galactic extinction as in Schlegel, Finkbeiner, & Davis
(1998), using the Fukugita et al. (2004) extinction law
in the SDSS bands and the Cardelli, Clayton, & Mathis
(1989) law in the UV. The fluxes were converted to mag-
nitudes on the AB system (Oke 1990) using zero point
fluxes of 3631 Jy in each band.
For the three galaxy pairs that span two SDSS fields
(Arp 85, 101, and 285), the optical images shown are the
mosaicked SDSS images from Hogg et al. (2007). How-
ever, the magnitudes quoted in Table 3 for Arp 101 and
285 were measured on the original unmosaicked images.
For Arp 270 and 297, we also had to use adjacent SDSS
images to do the SDSS photometry of the ends of the
tails. For Arp 85, the angular size is so large that sky
measurements on the SDSS images is problematic, thus
we do not provide SDSS magnitudes.
6.2. Disk vs. Tidal Colors, Compared to Population
Synthesis Models
In Figures 19 − 23, we plot various UV − optical
color-color plots for the main bodies and the tidal fea-
tures of the Arp galaxies.
plotted separately from their parent tidal features. Se-
lected features are labeled in these figures.
plots, we superimpose evolutionary tracks from ver-
sion 5.1 of the Starburst99 population synthesis code
(Leitherer et al. 1999), assuming instantaneous bursts,
Kroupa (2002) initial mass functions, and an initial mass
The candidate TDGs are
On these

Page 4

4Smith et al.
range of 0.1 − 100 M⊙. This version of the code in-
cludes the Padova asymptotic giant-branch stellar mod-
els (V´ azquez & Leitherer 2005).
lar metallicity and 0.2 solar metallicity models on these
plots. We convolved the model spectra with the SDSS
and GALEX filter response functions to obtain the model
colors shown in Figures 19 − 23.
As in our Arp 285 paper (Smith et al. 2008), to these
model colors we added in the Hα line, which can con-
tribute significantly to the r band flux for very young
ages. We note that this effect is redshift-dependent. For
the models plotted in Figures 19 − 23, for the Hα line we
assumed the velocity of Arp 285, which is typical of the
sample as a whole (Table 1). At this redshift, for very
young star forming regions (1 − 5 Myrs), the r magnitude
decreases by 1.1 − 0.25 magnitudes due to the presence of
Hα (Smith et al. 2008). However, for our highest redshift
objects, the Hα line is shifted to the edge of the r band fil-
ter where the sensitivity is down by a factor of ∼2.5. For
these high redshift galaxies, the model r magnitudes plot-
ted in Figures 19 − 23 for very young ages are too bright.
The effect of Hα is illustrated in Figure 21, the g − r vs.
u − g plot, in which we show tracks with and without
Hα. For the rest of the plots, we only display the models
which include Hα. Note that the plotted models do not
include the [O III] λ5007 or Hβ line, which can contribute
substantially to the g filter for low metallicity young
galaxies (e.g., Kr¨ uger, Fritze-Alvensleben, & Loose 1995;
West et al. 2009).
In Figures 24 − 28, we provide histograms of various
colors for the main bodies of the galaxies, the tidal fea-
tures, and the candidate TDGs separately. Selected fea-
tures are labeled in these figures. For the main bodies of
the interacting galaxies, there is a range of colors, reflect-
ing a range in star formation histories, extinction, and
progenitor morphological types. For example, the four
reddest disks in FUV − NUV are all apparently early-
type galaxies: Arp 173 N (classified as S0 in NED1), Arp
100 S = IC 19 (classified as E in NED), Arp 290 S = IC
195 (classified as SAB0 in NED), and Arp 120 N = NGC
4435 (classified as SB0 in NED). The UV colors of the
fifth reddest disk in FUV − NUV, Arp 89 W (= KPG
168 B), may be strongly affected by extinction, since it is
an edge-on disk galaxy classified as Sc in NED. For com-
parison, in the Nearby Galaxies Atlas (Gil de Paz et al.
2007), there is a correlation of FUV − NUV with mor-
phological type, with the early-type galaxies (E/S0) be-
ing redder, with FUV - NUV between 1 − 2, consistent
with our reddest systems.
In contrast, many of the bluest disks in our interacting
sample have very late morphological types according to
NED. For example, in NUV − g the bluest disks are
Arp 202 S (Im pec), Arp 24 main (SABm), and Arp 305
N (SBdm). These may have originally been late type
galaxies, or, alternatively, their morphology may have
changed due to the encounter.
We do not find a large difference in the colors of
the tidal features compared to those of the disks, on
average (see Figures 24 − 28).
that, on average, the tidal features do not have stronger
mass-normalized bursts of star formation than the disks.
We display both so-
These results suggest
1
TheNASA ExtragalacticDatabase;
http://nedwww.ipac.caltech.edu
Kolmogorov-Smirnov(K-S) tests cannot rule out the pos-
sibility that the colors of the disks, the tails/bridges, and
the TDGs come from the same parent population. The
most significant difference is found for g − r, where the
tidal features are slightly bluer than the disks, and a
K-S test gives a 2.5% probability that the two samples
come from the same population. This result is suggestive,
but inconclusive. Schombert, Wallin, & Struck-Marcell
(1990) also found possibly bluer B − V colors (approxi-
mately g − r) for the tidal features than the disks, while
their V − i colors (approximately r − i) for the main
bodies and tidal features were similar.
the Schombert, Wallin, & Struck-Marcell (1990) sample
contains more early-type galaxies than our sample. It
also contains a number of merger remnants, which were
excluded from our sample. The main bodies of these
galaxies are likely redder on average than our sample
galaxies due to older stars or more extinction.
Comparison of the colors of the tidal features with the
population synthesis models (Figures 19 − 23) show that,
in general, the light from both the disks and the tidal
features is not dominated by very young stars, and these
features are not completely extinction-free. As with the
main bodies of these galaxies, there is a range of colors
for the tidal features, due to a range in star formation
properties and original morphologies. For example, the
three tails that are reddest in FUV − NUV are the two
broad diffuse tails of Arp 283 and the southern tail in Arp
173, which extends from a red S0 galaxy. No clumps of
star formation are visible in these tails. In contrast to
these features, the bluest tail in FUV − NUV is Arp
120 WT, with FUV − NUV = −0.03. This feature was
originally discovered by Boselli et al. (2005) using these
same GALEX images.
We note that
6.3. The Colors of the Tidal Features vs. their Parent
Disks
Another useful test is to compare the colors of the in-
dividual tidal features with the colors of their own parent
disk, rather than with the sample as a whole. In Figure
29, in the various colors we provide histograms of the
differences between the colors of the disk and the col-
ors of their matching tidal features. For the bridges, we
matched with the most probable progenitor disk of the
two galaxies. We matched the northern Arp 285 tail and
the southern Arp 105 tail with the southern and north-
ern galaxies in those pairs, respectively, as those are the
most likely progenitors (see Section 7).
For most of the histograms in Figure 29, the distribu-
tions of color differences peak near zero, and the mean
color difference is smaller than the average measurement
uncertainty. For NUV − g, however, there is a significant
shift of the histogram to the right of zero, with the mean
(NUV − g)disk− (NUV − g)tidalcolor being 0.26, com-
pared to a typical measurement uncertainty in this color
of 0.03 magnitudes in Table 3. Thus the tidal features
are, on average, bluer in NUV − g than their own parent
disks. This confirms that, on average, tidal features are
indeed more prominent in the UV than in the optical,
compared to their own parent disks. For g − r, there is
a marginal effect, in that the mean (g − r)disk − (g −
r)tidalis 0.09, while the average g − r measurement un-
certainty per feature is 0.03. We note that the Figure 29
colors for the tidal features do not include the fluxes of

Page 5

GALEX UV Atlas of Interacting Galaxies5
the candidate tidal dwarf galaxies, which were measured
separately. As seen in Figure 25, for Arp 181, 202, and
305, the TDGs are considerably bluer in NUV − g than
their parent tidal features. The inclusion of these regions
as part of the tidal feature would increase the difference
in color from their associated disk.
Inspection of the population synthesis models plotted
in Figure 20 shows that a shift to the blue for NUV −
g without a strong change in g − r is likely an age ef-
fect rather than an extinction effect. With the GALEX
data, we are able to break the age-extinction degener-
acy to a certain extent. Thus we suggest that the stars
in the tidal features are younger, on average, than those
in their parent galaxies, and star formation, rather than
dust extinction, is the primary factor responsible for the
bluer NUV − g colors.
In addition to age, another factor that may be impor-
tant is metallicity. In general, lowering the metallicity
makes the NUV − g colors bluer (see Figure 20). Of all
the colors investigated in this study, NUV − g is the most
strongly affected by metallicity. Since tidal features tend
to be drawn from lower metallicity regions in the outer
disks, tails and bridges likely have lower abundances than
their parent galaxies on average. Thus the observed dif-
ference may be due in part to lower metallicities in the
tidal features. Only a handful of the tidal features in our
sample have available oxygen abundances, which range
from log(O/H) + 12 of 8.4 to 8.7, or ∼1/3− 1/2 solar (see
Section 7). More direct measurements of abundances in
our sample galaxies would be useful to better distinguish
between age and metallicity as the cause for this color
difference.
6.4. Colors of Subsets of the Sample
As noted in Section 5, our sample is comprised of pre-
merger interacting pairs of galaxies with a large variety
of morphologies. The sample includes both close pairs
and wide pairs, equal mass pairs as well as unequal mass
pairs, and pairs with long tails as well as pairs with short
or weak tails. To investigate whether any of these sub-
types of interacting galaxies stands out from the rest of
the sample in their UV/optical colors, we have selected
four different subsets from the sample, two selected based
on morphology, and two based on pair separation. First,
we divided the sample into two groups based on pair sep-
aration with the ‘wide pairs’ being the 15 pairs with sep-
arations greater than or equal to 30 kpc, and the ‘close
pairs’ sample being the 27 galaxies with separations less
than 30 kpc (see Table 1). Next, we created a subset
including the 10 M51-like galaxies in the sample (see Ta-
ble 1). Finally, we separated out another subset that
includes the 10 pairs with disturbed disks but with weak
or no tidal features visible in the optical/UV images. In
Figure 30, we present histograms of FUV − NUV, NUV
− g, and g − r colors of the galactic disks for these subsets
of galaxies. We used K-S tests to search for significant
differences in the distributions of optical/UV colors of
these subsets of galaxies. In no case were we able to rule
out the hypothesis that they came from the same parent
population.
These results are consistent with our earlier study
on the Spitzer colors of these galaxies, in which we
could not find significant differences between wide and
close pairs, and between M51-like systems and the rest
of the sample (Smith et al. 2007).
results for the SB&T galaxies, earlier studies based
on Hα equivalent widths and optical spectroscopy de-
tected a significant difference in star formation rate
between close and wide pairs, though with a large
amount of scatter (Barton et al. 2000; Lambas et al.
2003; Nikolic, Cullen, & Alexander 2004). Our different
conclusions are likely caused in part to selection effects.
Our pairs were selected based on the presence of strong
tidal distortion. Many of our widely-separated pairs have
very long tidal features, which indicates that they prob-
ably were closer together at some point in the past, and
have already undergone a strong gravitational encounter.
In contrast, in the earlier studies, pairs were selected
based solely on proximity in space, and thus many of
the wider pairs may not have experienced such strong
tidal disturbances. Another factor is simply sample size,
as our subsets of galaxies are quite small, compared to
these earlier studies. Larger sample sizes are needed to
further investigate whether particular classes of interact-
ing galaxies are more likely to have enhanced UV/optical
colors than other types.
In contrast to our
6.5. UV/Optical Colors vs. Spitzer IR Colors
Broadband colors in both the UV/optical range and
in the mid-infrared regime are sometimes used as in-
dicators of recent star formation in galaxies, however,
each method has advantages and disadvantages. Younger
stars have bluer UV/optical colors, however, dust extinc-
tion and/or metallicity differences can redden these col-
ors, confusing the issue. The Spitzer 3.6 µm broadband
filter is generally assumed to be dominated by the stellar
continuum from the older stellar population, while the 24
µm band is dominated by emission from hot dust heated
by UV photons from young stars, thus redder [3.6 µm] −
[24 µm] color are associated with younger stellar popula-
tions on average (e.g., Smith et al. 2007). This is also the
case for the [3.6 µm] − [8.0 µm] color, however, in addi-
tion to hot dust the Spitzer 8 µm band also includes the
prominent interstellar polycyclic aromatic hydrocarbon
(PAH) features, which can vary from galaxy to galaxy
depending upon the chemistry and the hardness of the
UV radiation field. In addition, in extreme starbursts, in-
terstellar contributions can be significant in the 3.6 µm
Spitzer band (Smith & Hancock 2009).
To determine how well the colors in these two wave-
length regimes correlate, in Figure 31 we compare
the Spitzer [3.6] − [24] and [3.6] − [8.0] colors for
the disks and tidal features in our sample with their
GALEX/SDSS FUV − NUV, NUV − g, and g − r colors.
The Spitzer magnitudes were obtained from Smith et al.
(2007) when possible. For the galaxies in our current
sample that are not in Smith et al. (2007), when Spitzer
images were available in the Spitzer archives, we down-
loaded these images and extracted magnitudes for the
appropriate regions.As noted earlier, in most cases,
the GALEX/SDSS magnitudes were measured over the
same area of the sky as the Smith et al. (2007) Spitzer
magnitudes. However, in the current study we measured
the GALEX/SDSS magnitudes of candidate TDGs sep-
arately from their host tidal feature; this was not done
for the Spitzer fluxes in Smith et al. (2007). Also, in a
few cases we modified the Spitzer regions somewhat. For
these galaxies, we re-determined the Spitzer fluxes in re-