Detection of CO(1−0) Emission at the Tips of the Tidal Tail in the Antennae Galaxies

The tip of the tidal tail, resulting from an encounter between galaxies, features gas concentrations and some star-forming regions, such as tidal dwarf galaxies (TDGs). This region provides a unique laboratory for examining the star formation process in a dynamical environment distinct from that of disk galaxies. Using the Nobeyama 45 m telescope, we conducted 12CO(1−0) position-switching observations at the tips of the southern tidal tail in the Antennae galaxies. We detected CO emission not only from the two star-forming TDG candidates but also in regions with no significant star formation. Adopting a Galactic CO-to-H2 conversion factor without helium correction, the H2 gas surface density is ∼5–12 M ⊙ pc−2. In most regions, the molecular-to-atomic gas ratio is around unity (0.6–1.9), but we find a region with a high ratio with a 3σ lower limit of >7.2. The star formation efficiency (SFE) of molecular gas is notably low (<0.15 Gyr−1), indicating less active star formation than in both nearby disk galaxies (∼0.5–1.0 Gyr−1) and other TDGs previously observed. Including previous observations, the molecular gas SFEs vary widely among TDGs/tidal tails, from 10−2 to 10 Gyr−1, demonstrating significant variations in star formation activity. Potential factors contributing to the low SFE in the Antennae tail tips include extensive tides and/or the young age of the tail.


Introduction
One of the major challenges in astrophysics involves understanding the process of star formation in galaxies.Specifically, it is crucial to understand how the star formation rate (SFR) correlates quantitatively with both atomic and molecular gas, and how this relationship varies depending on the environment.The relationship among the surface densities of SFR (Σ SFR ), H I gas (Σ H I ), and H 2 gas ( H 2 S ) is commonly referred to as the Kennicutt-Schmidt (K-S) relation (e.g., Kennicutt 1998).For nearby star-forming disk galaxies, the spatially resolved K-S relation on kiloparsec/subkiloparsec scales has been studied well (e.g., Kennicutt 1998;Bigiel et al. 2008Bigiel et al. , 2011;;Schruba et al. 2011;Sun et al. 2023).These studies showed no clear correlation between Σ SFR and Σ H I but a tight relationship between Σ SFR and H 2 S across a wide range of M 10 10 pc . Is this picture of star formation Universal?To reveal commonalities and differences in the star formation process across the Universe, it is important to expand the target to other types of galaxies, not only disk galaxies, and to investigate the star formation process in various dynamical environments.Indeed, it has been demonstrated that starburst galaxies, such as submillimeter-selected galaxies and ultraluminous infrared galaxies, follow a distinct sequence in the K-S relation compared to nearby disk galaxies (e.g., Komugi et al. 2005;Daddi et al. 2010;Genzel et al. 2010;Kennicutt & De Los Reyes 2021).Even within the disk galaxies, the K-S relation has been reported to tend to depend on the galactic environment such as bar, bar end, and center (e.g., Momose et al. 2010;Maeda et al. 2020Maeda et al. , 2023)).
Tidal dwarf galaxies (TDGs), which are self-gravitating lowmass galaxies formed from tidal debris of interacting galaxies (e.g., Zwicky 1956;Barnes & Hernquist 1992;Mirabel et al. 1992;Elmegreen et al. 1993), provide unique laboratories for examining the star formation process in a dynamical environment distinct from that of disk galaxies.This distinction arises first from the absence of large-scale gravitational fields such as spiral arms, which in disk galaxies play a crucial role in compressing and accumulating molecular gas to form massive stars (e.g., Egusa et al. 2011).TDGs do not possess such largescale gravitational fields.Moreover, TDGs are subject to tidal forces and likely to be free of nonbaryonic dark matter because tidal forces effectively segregate baryons in the disk from dark matter in the halo of the parent galaxies (e.g., Lelli et al. 2015).These features make TDGs a good comparison with nearby disk galaxies in which to study star formation in extreme dynamical environments.
As TDGs are recycled objects, they inherit their parent galaxies' metallicity, making CO a suitable molecular gas tracer for comparison with nearby disk galaxies (e.g., Duc et al. 2000).Previous CO observations toward TDGs have shown that molecular gas is abundant in TDGs (e.g., Braine et al. 2000Braine et al. , 2001;;Lisenfeld et al. 2002Lisenfeld et al. , 2004Lisenfeld et al. , 2008Lisenfeld et al. , 2016;;Duc et al. 2007;Boquien et al. 2011;Querejeta et al. 2021;Kovakkuni et al. 2023).Whether the molecular K-S relation (i.e., H 2 S versus Σ SFR ) of the TDGs is consistent with that in the disk galaxies is under debate.Braine et al. (2001)  = S S ) in the TDGs is comparable to that in spiral galaxies.On the other hand, TDGs that do not follow the molecular K-S relation of normal disk galaxies have been found (Lisenfeld et al. 2008(Lisenfeld et al. , 2016;;Querejeta et al. 2021;Kovakkuni et al. 2023), suggesting a diversity of SFE H 2 in TDGs.In particular, Querejeta et al. (2021) reported that there are regions in tails where star formation is severely suppressed despite the presence of large amounts of molecular gas.This result indicates the need for unbiased CO observations that focus not only on star-forming regions but also on other regions where no star formation is seen.
In this paper, we report 12 CO(1−0) observations of the tips of the southern tidal tail in the Antennae galaxies (NGC 4038/ 39, Arp 244; Figure 1) with the 45 m telescope of the Nobeyama Radio Observatory (NRO).The Antennae galaxies are well-known nearby interacting galaxies (22 Mpc;Schweizer et al. 2008) with long H I gas-rich tails (Hibbard et al. 2001).These galaxies showcase the intense tidal interactions between disk galaxies of similar mass (e.g., Toomre & Toomre 1972), with simulations suggesting the collision occurred 100-300 Myr ago (Renaud et al. 2015).It is clear that the H I gas is concentrated at the tip where the H I column densities per unit velocity are higher than anywhere else in the system, including within the main disks (Hibbard et al. 2001).Two TDG candidates (black contours in Figure 1(b)) at the tips of the southern tidal tail, discovered by Schweizer (1978) and Mirabel et al. (1992), are suggested to be young and formed in situ (Saviane et al. 2004;Hibbard et al. 2005), but whether they are self-gravitating systems or not is not clear (Hibbard et al. 2001).The metallicity of 12 log O H 8.4 et al. 1992) is about half-solar.Interestingly, no star formation is seen in the regions north of the two TDG candidates (i.e., regions with decl. 19 00 <- ¢) despite H I being abundant.Targeting this region as well as the two TDG candidates, we performed position-switching CO observations to examine the K-S relations.New CO observations and ancillary data are described in Section 2. Results on molecular gas contents and the K-S relation in the Antennae tidal tail tips are presented in Section 3. In Section 4, we discuss the star formation activity there.Section 5 gives a summary of this paper.The velocity refers to LSR velocity.
and the size of each beam is ∼14″ at 115 GHz, corresponding to 1.5 kpc at the Antennae system.The efficiency of the main beam reported by the observatory during 2022-2023 season was η mb = 0.39 ± 0.03.
We observed 12 positions across three setups, labeled A-1 to C-4 in Figure 1 The backend is an FX-type correlator system, SAM45, comprising 16 arrays with 4096 spectral channels each.Two arrays are allocated for each beam and polarization (i.e., the same polarization data are input to the two arrays).The bandwidth and resolution were set at 2 GHz and 488.28 kHz, respectively, corresponding to 5200 and 1.3 km s −1 at 115 GHz.We employed the position-switching mode, with an on-source integration time of 10 s per scan.The off point was set to be 7 5 south of the observation position.Telescope pointing was checked every 50-60 minutes by observing SiO masers close to the targets.The typical pointing error was under 3″.The line intensity was calibrated by the chopper wheel method.The system temperature (T sys ) was 350-550 K, with the elevated T sys largely attributed to the low elevation of 25°−35°.
The observed data were analyzed as follows.First, we flagged the scan taken with a wind velocity larger than 5 m s −1 .Next, we subtracted a baseline (second-order polynomial function) from each scan and flagged any scan with poor baselines or/and spurious lines manually.Then, for each array, we summed the scans to create a spectrum.For two arrays with the same polarization input, the array exhibiting the bestsubtracted baseline was chosen.The two polarizations were then merged to generate the final spectrum.The effective onsource integration time after flagging for each position is ∼0.3-0.5 hr (Table 1).We converted the antenna temperature (T A *) into the brightness temperature of the main beam (T mb ) using T T . Finally, we smoothed the spectrum by binning to 10 km s −1 .The resultant rms noise (T rms ) ranged from 14 to 18 mK (Table 1).

Ancillary Data
H I data.We used H I data obtained from the Very Large Array (VLA), originally published by Hibbard et al. (2001).The data were taken by the VLA C+D array with an angular resolution of 20 7 × 15 4 and velocity resolution 5 km s −1 .The noise level is 0.9 mJy beam −1 , corresponding to 1.6 × 10 −19 cm −2 beam −1 channel −1 .We smoothed the data cube to a 10 km s −1 bin.
GALEX far-ultraviolet (FUV) and Spitzer 24 μm.As SFR tracers, we used GALEX FUV and Spitzer 24 μm archival data. 14For the GALEX FUV image, we use the delivered images from the z = 0 Multiwavelength Galaxy Synthesis GALEX-WISE Atlas data release 1 (Leroy et al. 2019).We utilized the background-subtracted FUV image that has a Gaussian point-spread function (PSF) with an FWHM of 15″.This delivered image had already been corrected for Galactic foreground extinction of A FUV = 0.41 mag.
We downloaded the Spitzer 24 μm image, which was observed with the Multiband Imaging Photometer, from the Spitzer Heritage Archive.The image was processed with the Super-Mosaic Pipeline.We subtracted the average sky background determined as the mode value of blank sky.Then, we convolved the image to an image that has the same PSF as the GALEX image.Here, we used the kernel delivered by Aniano et al. (2011).
(2) The effective on-source integration time.
(3) The 1σ rms in 10 km s −1 bins.(4) The peak brightness temperature of the CO(1−0) emission.( 5), (6) We do not include the contribution from helium.(9) The unobscured SFR surface density from GALEX FUV.(10) The upper limit of total SFR surface density, which is the value in column (9) plus a 3σ upper limit of the embedded SFR surface density from Spitzer 24 μm.
with the velocity ranges of detected emissions highlighted in yellow regions.Our result shows that molecular gas is distributed over a wide area (at least ∼15 kpc) at the tip of the southern tail of the Antennae galaxies: we detected CO emissions not only in the two TDG candidates (B-1 and C-2 in TDG S78; B-4, C-3, and C-4 in TDG MDL92) but also in the regions where significant star formation is not seen (i.e., A-2, A-3, and A-4).We derived the velocity-integrated intensities of CO(1−0) (I CO(1−0) ) from the CO emission shown as the yellow regions in Figure 2. H 2 S is derived from I CO(1−0) as where H 2 S is in units of M e pc −2 and I CO(1−0) is in units of K km s −1 .We adopt the Galactic CO-to-H 2 conversion factor of 3.20 M K km s pc (e.g., Bolatto et al. 2013) as with the previous TDG observations (e.g., Braine et al. 2001;Querejeta et al. 2021;Kovakkuni et al. 2023).This conversion factor does not include any contribution from helium.To scale our quoted surface densities to account for helium, they would need to be multiplied by a factor of ∼1.36.Since the metallicity of the TDG candidates is about half-solar (Mirabel et al. 1992), the CO-to-H 2 conversion factor may be larger than the Galactic one and H 2 S may be underestimated by a factor of 2-3.However, this underestimation does not change the discussion and conclusions of this paper.We estimate a 3σ upper limit for the CO(1−0) flux in B-2 and B-3 as , where ΔV ch is the channel width of 10 km s −1 and ΔV line is the assumed line width based on the H I line.
The resultant I CO(1−0) and H 2 S are summarized in Table 1.
S is in the range ∼5-12 M e pc −2 , corresponding to a molecular gas mass in the beam size of ∼(0.7-2.1)× 10 7 M e .
CO was detected at positions A-1, A-4, C-3, and C-2, which are near the undetected positions B-2 and B-3.This suggests that the distribution of molecular gas in the Antennae tidal tail tips is somewhat clumped, as noted in the recent CO mapping toward other TDGs (Querejeta et al. 2021).Future CO mapping observations will reveal the molecular gas distribution.When compared with the previous IRAM 30 m CO(2-1) observations toward B-4 by Braine et al. (2001, see their Figure 4), the velocities at the emission peaks are consistent at ∼1690 km s −1 .The ratio of peak temperature is CO(2-1)/CO (1−0) ∼ 0.55.This value is comparable to those reported in other TDGs of 0.56 (Lisenfeld et al. 2002) and 0.5 (Querejeta et al. 2021).Moreover, these CO(2-1)/CO(1−0) ratios are comparable to those in nearby disk galaxies at kiloparsec scales, typically ranging from 0.4 to 0.9, with a median of 0.61 (Yajima et al. 2021).

Comparison of Molecular Gas and Atomic Gas
In Figure 2, red dashed-dotted lines show the H I profiles.We derived H I S from 21 cm line intensities (I H I ) as where H I S is in units of M e pc −2 and I H I is in units of K km s −1 (e.g., Schruba et al. 2011).We assume that the H I is optically thin.Like H 2 S , H I S does not include the contribution from helium.In the same way as in Section 3.1, the 3σ upper limit at A-4 is derived based on the CO emission.H I S is ∼2-9 M e pc −2 (Table 1).At most positions, the velocities of the CO and H I emission peaks are generally consistent, with a difference of no more than 20 km s −1 .Such coincidences of the peaks have been reported in many TDGs (e.g., Braine et al. 2000Braine et al. , 2001)) S S has been also reported in regions with no star formation in TDG J1023+1952 (Querejeta et al. 2021).

Kennicut-Schmidt Relation
In our measurements, Σ SFR is calculated from a linear combination of GALEX FUV and Spitzer 24 μm intensities by Leroy et al. (2008) as S m are unobscured and embedded SFR terms with units of M e yr −1 kpc −2 , respectively, and I FUV and I 24μm are the FUV and 24 μm intensities in units of MJy sr −1 , respectively.These equations assume the default initial mass function of STARBURST99 (Leitherer et al. 1999) and the broken power low by Kroupa (2001)  S m (2.8 × 10 −4 M e yr −1 kpc −2 ) is much higher than that of SFR FUV

S
(3.0 × 10 −5 M e yr −1 kpc −2 ), embedded SFR components that are reemitted as 24 μm are likely to be under the noise level.Thus, the true SFR would be between SFR FUV

S
and the upper limit of Σ SFR , which are listed in Table 1.
Figure 3(a) shows the molecular K-S relation for Antennae tidal tail tips compared to nearby disk galaxies and TDGs from the literature.The orange, red, and cyan symbols show our measurements at the positions A-1 to C-4.The filled and open symbols represent the unobscured SFR and the upper limit of the total SFR, respectively.The gray circles in the background show the kiloparsec-scale measurements in nearby disk galaxies from the PHANGS-Atacama Large Millimeter/ submillimeter Array (ALMA) project (Sun et al. 2022) = S S , in the star-forming regions in Antennae tidal tail tips is notably low (<0.15Gyr −1 ) relative to that in nearby disk galaxies (∼0.5-1.0Gyr −1 ), indicating inactive star formation compared to the normal disk region.
Figure 3(a) also displays the molecular K-S relation in TDGs from the literature as colored symbols.While earlier IRAM 30 m observations toward star-forming regions in TDGs reported that TDGs follow the standard K-S relation seen in nearby disk galaxies (Braine et al. 2001;Lisenfeld et al. 2002;Boquien et al. 2011), more recent studies, including ours, highlight the diversity of star formation activity in TDGs (Lisenfeld et al. 2016;Querejeta et al. 2021;Kovakkuni et al. 2023).While some TDGs with low SFEs similar to our results have been reported (VCC 2062, Lisenfeld et al. 2016;J10231952, Querejeta et al. 2021), there are TDGs reported to be in the starburst regime (i.e., SFE 10 Gyr H 1 2 ~-) of the K-S relations (NGC 5291N and NGC 5291S, Kovakkuni et al. 2023).Note that Kovakkuni et al. (2023) likely overestimated the SFE because their ALMA 12 m array observations miss some CO flux on larger scales.To summarize these observations, SFE H 2 differs by up to three orders of magnitude from 10 −2 to 10 Gyr −1 , suggesting intrinsic differences in the star formation process of individual TDGs/tidal tails.The causes of these SFE variations are discussed in Section 4. S S < ).This indicates a nonuniform relationship between total gas and SFR, as reported by previous studies (e.g., Kennicutt 1998;Bigiel et al. 2008;Schruba et al. 2011).This trend is well parameterized by a fixed SFE H 2 as suggested by Schruba et al. (2011).H H 2 I S S in the Antennae tidal tail tips is comparable to that in nearby disk galaxies.However, SFE tot is extremely low, causing these data points to deviate significantly from the relationship observed in nearby disk galaxies.
As seen in Figure 3(a), Figure 3(b) also shows diversity in the relationship between H H 2 I S S and SFE tot for the TDGs.Our measurements are comparable to those in the northern part of J1023+1952 (Querejeta et al. 2021) and VCC 2062 (Lisenfeld et al. 2016).It should be noted that some of the literature values would be unrealistically low surface densities in Figure 3, caused by beam dilution due to poor angular resolutions.The spatial resolution of CO images ranges from 600 pc (NGC 5291N and S, Kovakkuni et al. 2023) to 12.2 kpc (Arp 105S, Braine et al. 2001).Comparisons between TDGs at the same spatial resolution will be needed in the future.

Discussion
Our CO observations reveal the presence of molecular gas in regions with no apparent star formation activity (A-2, A-3, and A-4), where H 2 S is comparable to that in the star-forming regions.This indicates that the absence of star formation is not attributable to a lack of molecular gas.The type of tidal force, which varies between compressive and extensive depending on the shape of the gravitational potential, is likely to be a critical factor.The distribution of these two modes would play a role in the formation and evolution of TDGs (e.g., Renaud et al. 2009Renaud et al. , 2015;;Ploeckinger et al. 2015).In this mechanism, a fully compressive mode of tidal forces triggers the formation of TDGs and massive star clusters.Therefore, the star-forming regions of TDG candidates in the Antennae tidal tail tips might be regions of compressive tides, while the regions to the north of them without star formation are thought to be regions of extensive tides.Simulations modeling the Antenna galaxies reproduced this picture (Renaud et al. 2015, see their Figure 19).Such extensive tides are expected to prevent the formation of dense and compact components, i.e., giant molecular clouds (GMCs).According to the simulation by Renaud et al. (2015), in regions of extensive tides, molecular gases tend to exist as diffuse extended components, which would not directly contribute to the current star formation activity.It is shown that such a large amount of diffuse molecular gas makes SFE low in a nearby barred galaxy (Maeda et al. 2020).In fact, a high fraction of diffuse molecular gas has been reported in TDG J1023+1952, which exhibits an extremely low SFE like our results (Figure 3).It is shown that GMCs contribute only 10%-20% of the total CO flux in that region, while the remaining 80%-90% is attributed to extended diffuse molecular gas distributed on a scale of 2 kpc or more (Querejeta et al. 2021).
Even though star-forming regions are seen in the two TDG candidates, their Σ SFR are significantly lower than expected from the K-S relation for nearby disk galaxies (Figure 3).This may be due to weak compressive tides.According to the

S
= S + S m , respectively.Other colored symbols refer to Braine et al. (2001, B+01), Lisenfeld et al. (2002, L+02), Boquien et al. (2011, B+11), Lisenfeld et al. (2016, L+16), Querejeta et al. (2021, Q+21), and Kovakkuni et al. (2023, K+23).The beam sizes in parsecs of these observations are displayed in the legend column and "mean" indicates the mean value in the entire CO-emitting area.For Querejeta et al. (2021), measurements for the north (N) and south (S) TDG parts are displayed.The gray circles in the background show the kiloparsec-scale measurements in nearby disk galaxies from the PHANGS-ALMA project (Sun et al. 2022).(b) Same as panel (a), but focusing on the molecularto-atomic gas ratio vs. the total gas SFE.The dotted lines in each panel represent constant molecular gas SFEs.
simulation by Renaud et al. (2015), in the Antennae southern tail, compression tides are weak and exist as several small regions with a size of <500 pc, suggesting that extensive star formation on scales larger than 500 pc is unlikely to occur.Therefore, in the beam of Nobeyama 45 m (∼1.5 kpc), a significant amount of diffuse gas is thought to exist in the starforming regions, leading to low SFE.
Furthermore, we note that recent numerical simulations studying the internal structure of individual GMCs suggest that the rotational mode of turbulence becomes dominant in clouds to decrease SFE if the clouds are formed through the compression of highly inhomogeneous interstellar medium (e.g., Kobayashi et al. 2022).This means that SFE may vary even in compressive tides depending on the degree of density inhomogeneity in the tides.
In another scenario, the current low SFE might be attributed to the youth of the TDGs, capturing them in stages before star formation.Duc et al. (2004) proposed a slightly different scenario than the compressional tidal scenario: accumulation and collapse of massive gaseous condensations at the end of the tidal tails.In this scenario, the tidal field efficiently transports a substantial amount of gas from the disk, preserving its surface density at a high level, leading to density enhancements at the tidal tail tips.Eventually, self-gravity dominates, resulting in the collapse of gas clouds and the onset of star formation.
The star-forming TDGs NGC 5291N, NGC 5291S, and NGC 7252SW are associated with significant H I concentrations and exhibit a velocity gradient that is decoupled from the surrounding tidal debris, indicating rotation within a selfgravitating potential well (Bournaud et al. 2007;Lelli et al. 2015).However, for the two TDG candidates in our study, H I observations did not reveal any kinematic signatures of selfgravitating entities (Hibbard et al. 2001).(Note that we cannot rule out the possibility that such a signature is masked by the bending of the tip of the tail.)According to simulations, NGC 5291 and NGC 7252 collided 360 Myr ago (Bournaud et al. 2007) and 700 Myr ago (Hibbard & Mihos 1995;Chien & Barnes 2010), respectively, while the Antenna galaxy collided 100-300 Myr ago (Renaud et al. 2015).Therefore, the TDG candidates at the Antennae tidal tail tips are relatively young and we may be observing the stage before the gas collapses into a self-gravitating system and extensive star formation begins.
To uncover the reasons behind the notably low SFE at the tips of the Antennae tails, it is crucial to conduct highresolution CO mapping using interferometry, as well as CO mapping with a single dish to reveal total CO flux.Assessing the fraction of diffuse gas, examining the GMC properties, and analyzing the velocity field of the molecular gases will provide valuable insights into this phenomenon.

Summary
Using the Nobeyama 45 m telescope, we conducted 12 CO(1−0) position-switching observations (beam size of 14 1.5 kpc ¢ = ¢ ) at the tips of the southern tidal tail in the Antennae galaxies to examine the K-S relations (Figure 1).As ancillary data, we use H I data from VLA, and GALEX FUV and Spitzer 24 μm as SFR tracers.We detected CO(1−0) emission lines from 10 out of 12 observed positions (Figure 2 and Table 1).We detected CO emissions not only at the two TDG candidates but also in regions with no significant star formation.Adopting a Galactic CO-to-H 2 conversion factor without helium correction, H 2 S is ∼5-12 M e pc −2 .In most regions, H HI 2 S S is around unity (0.6-1.9), but we find a region with a high ratio ( 7.2 H HI 2 S S > ).SFE H 2 is notably low (<0.15Gyr −1 ; Figure 3(a)), indicating less active star formation than in both nearby disk galaxies (∼0.5-1.0Gyr −1 ) and other TDGs previously observed.Including previous observations, the molecular gas SFEs vary widely among TDGs/tidal tails, from 10 −2 to 10 Gyr −1 , demonstrating significant variations in star formation activity.SFE tot is also extremely low compared to that in nearby disk galaxies (Figure 3(b)).One of the factors contributing to the SFE in the Antennae tail tips is the influence of extensive tides.Due to these extensive tides, GMCs may be less likely to form, resulting in molecular gas existing as diffuse extended components.Another potential reason is that the TDG candidates are relatively young.We might be observing a phase prior to the gas collapsing into a self-gravitating system and initiating extensive star formation.

Figure 1 .
Figure 1.(a) Overlay of DSS blue image and H I emission contours (gray; Hibbard et al. 2001), equivalent to Σ H I = 1, 3, 5, 10, 20 M e pc −2 .(b) Overlay of the H I image and SFR contours derived from the Galaxy Evolution Explorer (GALEX) far-ultraviolet (FUV) image (white), equivalent to M 0.5, 1.0, 5.0 10 yr kpc SFR FUV 4 1 2 ( ) S = ´--- , in the rectangular region in panel (a).The filled black circle in the lower left corner represents the beam size of the H I image.The orange, red, and cyan circles represent the observed positions with the Nobeyama 45 m telescope and the circle size shows the beam size of 14″.Circles of the same color indicate four beams that can be simultaneously observed with the Nobeyama 45 m telescope, linked together in a square by auxiliary lines for clarity.Thick black contours correspond to Σ H I = 6 M e pc −2 and delineate the approximate half-light level around the TDG candidates identified byMirabel et al. (1992, labeled MDL92)  and bySchweizer (1978, labeled S78).The labeling of these TDGs followsHibbard et al. (2001).
(b).A, B, and C show each setup and the numbers from 1 to 4 correspond to beams 1 through 4 of FOREST.The center coordinates of the four beams of the setup A, B, and C are (12 h 01 m 25 98, −18°59′24 0), (12 h 01 m 23 84, −19°00′09.″0),and (12 h 01 m 23 84, −19°00′34.″0),respectively.The beam positions cover a broad region at the tips of the Antennae southern tail, allowing us to compare where star formation is seen and not seen (see the contours of Σ SFR in Figure 1(b)).Position B-4 corresponds to the regions where previous CO(2-1) observations were conducted using the IRAM 30 m telescope (Braine et al. 2001).

Figure 2 .
Figure 2. CO(1−0) emission profiles (black lines) at the tips of the tidal tail in the Antennae galaxies obtained with the Nobeyama 45 m telescope.The observed positions are shown in Figure 1(b) with circles.The velocity ranges with >0.5T rms were defined as detected emissions and are highlighted in the yellow regions.The red dashed-dotted lines show the H I profiles obtained from the VLA(Hibbard et al. 2001).We show the profiles in the range ±250 km s −1 from the CO intensityweighted mean velocities.

)
with a maximum mass of 120 M e .While significant FUV emissions ( are seen around the TDG candidates (Figure 1(b)), no signal is seen in the Spitzer 24 μm image.Because the 3σ upper limit of SFR 24 m

Figure
as a function of the molecular-to-atomic ratio ( H H 2 IS S).In the nearby disk galaxies, although SFE tot remains relatively constant in regions where molecular gas dominates (

Figure 3 .
Figure 3. (a) Molecular Kennicutt-Schmidt relation for Antennae tidal tail tips vs. TDGs and nearby disk galaxies from the literature.The orange, red, and cyan symbols show our measurements at the positions A-1 to C-4.The filled and open symbols represent SFR FUVSand SFR detected CO lines in eight TDGs with active star formation and reported that the star formation efficiency (SFE H

Table 1
Properties of Gas and Star Formation in the Antennae Tidal Tail Tips