HI mapping of the Leo Triplet: Morphologies and kinematics of tails and bridges

A&A 658, A25 (2022) https://doi.org/10.1051/0004-6361/202141565 c G. Wu et al. 2022 Astronomy & Astrophysics Himapping of the Leo Triplet Morphologies and kinematics of tails and bridges Gang Wu1,2 , David Martínez-Delgado3,?, Christian Henkel1,4,2, Pavel Kroupa5,6, Fabian Walter7, Nico Krieger7, Alberto D. Bolatto8, Timothy Robishaw9, Joshua D. Simon10, Álvaro Ibáñez Pérez11, Karl M. Menten1, and Jarken Esimbek2 1Max-Planck-Institut für Radioastronomie, Auf dem Hügel 69, 53121 Bonn, Germany e-mail: [email protected] 2Xinjiang Astronomical Observatory, Chinese Academy of Sciences, 830011 Urumqi, Xinjiang, PR China 3Instituto de Astrofísica de Andalucía, CSIC, 18080 Granada, Spain 4Astronomy Department, King Abdulaziz University, PO Box 80203, Jeddah 21589, Saudi Arabia 5Helmholtz-Institut für Strahlen und Kernphysik (HISKP), University of Bonn, Nussallee 14-16, 53115 Bonn, Germany 6Charles University in Prague, Faculty of Mathematics and Physics, Astronomical Institute, V Holešoviˇ ckách 2, 18000 Praha, Czech Republic 7Max Planck Institute for Astronomy, Königstuhl 17, 69117 Heidelberg, Germany 8Department of Astronomy, University of Maryland, College Park, MD 20742, USA 9National Research Council Canada, Herzberg Programs in Astronomy and Astrophysics, Dominion Radio Astrophysical Observatory, PO Box 248, Penticton, BC V2A 6J9, Canada 10 Observatories of the Carnegie Institution for Science, 813 Santa Barbara Street, Pasadena, CA 91101, USA 11 Asociación Astronómica AstroHenares, 28823 Coslada, Madrid, Spain Received 16 June 2021 /Accepted 22 October 2021 ABSTRACT A fully sampled and hitherto highest resolution and sensitivity observation of neutral hydrogen (Hi) in the Leo Triplet (NGC 3628, M 65/NGC 3623, and M 66/NGC 3627) reveals six Histructures beyond the three galaxies. We present detailed results of the morphologies and kinematics of these structures, which can be used for future simulations. In particular, we detect a two-arm structure in the plume of NGC3628 for the first time, which can be explained by a tidal interaction model. The optical counterpart of the plume is mainly associated with the southern arm. The connecting part (base) of the plume (directed eastward) with NGC3628 is located at the blueshifted (western) side of NGC 3628. Two bases appear to be associated with the two arms of the plume. A clump with a reversed velocity gradient (relative to the velocity gradient of M66) and a newly detected tail, that is to say M66SE, is found in the southeast of M 66. We suspect that M 66SE represents gas from NGC3628, which was captured by M 66 in the recent interaction between the two galaxies. Meanwhile gas is falling toward M 66, resulting in features previously observed in the southeastern part of M 66, such as large line widths and double peaks. An upside-down “Y”-shaped Higas component (M65S) is detected in the south of M 65, which suggests that M 65 may also have been involved in the interaction. We strongly encourage modern hydrodynamical simulations of this interacting group of galaxies to reveal the origin of the gaseous debris surrounding all three galaxies. Key words. galaxies: individual: Leo Triplet – galaxies: interactions – galaxies: ISM – galaxies: peculiar – radio lines: galaxies 1. Introduction Interactions between galaxies have significant impacts on their participants, leading to asymmetries, warps, and exchange of gas and momentum. These interactions can finally result in noncircular potentials, leading to an inflow of gas, enhancement of bars, and also triggering starburst activity in the galaxies. An excellent example is the Leo Triplet galaxy system (see the left panel of Fig. 1), also known as Arp317 (Arp 1966), mainly including the SAB(rs)a galaxy M65 (NGC 3623), the SAB(s)b galaxy M66 (NGC 3627), and the SAb pec galaxy NGC 3628. The most famous feature in this system is probably the spectacular tail (the plume) from NGC 3628 extending toward the east. The plume was first detected by Zwicky (1956) and Kormendy & Bahcall (1974) via optical observations. Then Rots (1978) and Haynes et al. (1979) reported neutral hydrogen (Hi) ?Talentia Senior Fellow. observations in the Leo Triplet and revealed a ≈150 kpc (rescaled to a distance of 11.3 ±0.5 Mpc) long Histructure1consistent with its optical counterpart. In these Hiobservations, an additional bridge-like structure is also detected extruding from NGC 3628 and likely pointing southward to M66. Rots (1978) applied the methods of Toomre & Toomre (1972) and basically recovered the morphologies and kinematics of Hiplume and bridge by assuming a recent (8 ×108years ago) tidal encounter between NGC3628 and M 66. However, there remained some tension between the model and the observational data as argued 1Anand et al. (2021) recently compiled a catalog of the best available distances of 118 galaxies. Here we adopt the average of the red giant branch distance of M 65 (11.3 ±1.1 Mpc) and the group distances (from the galaxy group and numerical modeling of their orbits) of M66 (11.32 ±0.48 Mpc) and NGC 3628 (11.3 ±1.1 Mpc) in Anand et al. (2021) as the distance to the Leo Triplet. See Jacobs et al. (2009) and Anand et al. (2021) for more details. Open Access article, published by EDP Sciences, under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Open Access funding provided by Max Planck Society. A25, page 1 of 18 A&A 658, A25 (2022) by Haynes et al. (1979). Moreover, according to the scenarios simulated by Toomre & Toomre (1972), the bridge seems to be too prominent for two galaxies with similar total masses (e.g., Rots 1978). In spite of that, major mergers (involving two galaxies with comparable mass) are the commonly cited mechanism to explain the observations. An alternative is that a nonmerger tidal interaction between passing galaxies with comparable masses pulls out tidal tails. Such strong encounters are only possible if dynamical friction on the extensive dark matter halos does not operate (Kroupa 2015;Renaud et al. 2016). Evidence for interactions either related to M66 or NGC 3628 have also been found in optical, radio continuum, spectroscopic, and polarization observations (e.g., Rots 1978;Haynes et al. 1979;Young et al. 1983;Baan & Goss 1992;Zhang et al. 1993;Wilding et al. 1993;Reuter et al. 1996;Chromey et al. 1998;Soida et al. 2001; Chemin et al. 2003;Dumke et al. 2011;We˙zgowiec et al. 2012). However, previous simulations and observations are mainly focused on the interaction between M 66 and NGC 3628. The role of the third member, M 65, in this triplet system has remained unclear. On the one hand, M 65 looks quiescent and no clear distortion is seen in it, which supports the view that M 65 is not involved in the interaction (e.g., Afanasiev & Sil’chenko 2005;Duan 2006). On the other hand, simulations of the interaction solely between NGC 3628 and M 66 still show some discrepancies with observations (e.g., Haynes et al. 1979; Jennings et al. 2015). An alternative scenario for the formation of the giant tidal tail associated with NGC 3628 has also been discussed more recently, the minor merger (i.e., a merger between a large and a less massive satellite galaxy). Within the hierarchical galaxy formation framework, minor mergers are expected to be significantly more common than mergers with comparable masses at the current epoch (e.g., Cole et al. 2000). If the satellite galaxies become tidally disrupted, they should leave behind extended low surface brightness substructures known as tidal streams (e.g., Martínez-Delgado et al. 2010,2015). Two typical examples in our local volume are the Sagittarius tidal stream surrounding our Milky Way (e.g., Majewski et al. 2003) and the Great Southern stream around the Andromeda galaxy (e.g., Ibata et al. 2001). Recently, Jennings et al. (2015) identified and characterized a compact massive star cluster, NGC 3628-UCD1, embedded in the base of the plume. It shows similar characteristics as the most massive Milky Way globular cluster, ωCentauri, which is believed to be a stripped dwarf galaxy remnant (e.g., Bekki & Tsujimoto 2019). NGC3628-UCD1 (see Sect. 4.1 below) is also surrounded by a resolved stellar population showing a typical S-shaped structure, which strengthens the scenario of a young tidal stream. Therefore, Jennings et al. (2015) suggested the possibility of a minor merger between NGC3628 and a dwarf elliptical galaxy (the progenitor of the star cluster NGC 3628-UCD1) adding additional complexity to the standard interaction scenario. The appendages, that is tails and bridges, provide a fossil record of the interaction history and pose constraints on dynamical models. To recover the morphologies and kinematics of these appendages provides a principal benchmark for dynamical models (e.g., Rots 1978;Haynes et al. 1979). Hiobservations bring enormous convenience to identifying appendages. For example, the Hiplume (Haynes et al. 1979) is more remarkable than the faint optical plume (Kormendy & Bahcall 1974) and so far no CO has been detected from the plume (Young et al. 1983). Meanwhile, to obtain a full view of the appendages, the observations should include the three galaxies but should also cover a more extended region. Previous Hiobservations covering the entire Leo Triplet system were all conducted with the Arecibo 305 m telescope with a resolution of about 40(Haynes et al. 1979;Stierwalt et al. 2009). Higher resolution and sensitivity observations are required to further constrain the dynamical model and to also clarify the origin of some interesting features observed in this system. For example, the Hiemission observed along the plume surprisingly shows an almost constant velocity field but contains two velocity regimes (Haynes et al. 1979). A “peculiar velocity” clump is detected in the southeast of M66 which is found to be not corotating with M66 (Haynes et al. 1979;Zhang et al. 1993). In this paper, we report high resolution and sensitivity fully sampled neutral hydrogen observations of the entire Leo Triplet system by combining VLA and Arecibo observations. We provide detailed morphologies and kinematics of the appendages and seek to understand the interaction history of this system. 2. Observations 2.1. VLA and Arecibo observations The VLA observations (Project code: AB1074, PI: A. Bolatto) were conducted on 2003 March 29th and 30th in the D configuration. Eighteen fields were used to completely cover the Leo Triplet and the integration time on each field is &40 min. The details of the observations are summarized in Table 1. Calibration of the VLA observations was performed with the VLA pipeline in AIPS2and the continuum was subtracted from the visibilities using the task “uvlin” with a linear fit. The visibilities are then imaged and cleaned in CASA (McMullin et al. 2007) using a “multiscale” clean in the “tclean” task. Within a clean mask defined by the “auto-multithresh” algorithm, the emission is cleaned down to 1.5mJy beam−1(≈3σ). The synthesized beam is 6100 .06 ×5900 .40 at PA ≈−65◦and the pixel size is 1500. The channel spacing is ∆V=20.7 km s−1(see Table 1). The rms noise in a single channel is σrms .0.43 mJy beam−1. The Arecibo observations (Project code: A1581, PI: Bolatto) were conducted over 11 days from 2002 April 7 through 2002 April 20. A ≈2◦ .2×1◦ .5 area of the sky was covered by the Arecibo observations using the on-the-fly mapping technique (see Fig. B.1) with a total observing time of about 27.25h. These data were used to fill the zero spacing of the VLA HI observations. The L-band Narrow dual-polarization frontend (Tsys ≈ 30 K) was used in combination with the correlation backend configured for 9-level sampling over a 6.125 MHz bandwidth with 2048 channels. This yielded a 1320 km s−1bandpass sampled in 0.64 km s−1-wide channels. At this frequency, the telescope has a full width at half maximum (FWHM) beam width of ≈3.30. The orthogonal polarizations were combined by weighting by the reciprocal of the squares of their noise levels. The polarization-averaged on-the-fly (OTF) data were sampled onto a regular grid after applying a slant orthographic projection and a Gaussian convolution function with an FWHM of 30 .3. An effective integrated time per beam is about 70 s. The rms noise in a single channel is σrms .3.6 mJy beam−1. The interferometric and single dish cubes were combined with “immerge” (revision 1.7) in miriad (Sault et al. 1995) and the weighting factor between Arecibo and VLA is set to unity. The rms noise in a single channel of the combined data is σrms .0.5 mJy beam−1. Integrated over the full velocity range of a single channel, this yields σch =σrms ×∆V.0.01 Jy beam−1km s−1. 2http://www.aips.nrao.edu/ A25, page 2 of 18 G. Wu et al.: Himapping of the Leo Triplet Table 1. Summary of the VLA observations. Obs. date Array Freq. cove. (MHz) Spec. res. Bandpass Gain Flux configuration spw 1 spw 2 kHz (kms−1) cal. cal. cal. (1) (2) (3) (4) (5) (6) (7) (8) 2003 March 29–30 D 1416.198–1419.226 1413.854–1416.882 97.665 (20.7) 3C 147 1120+143 3C 286 Notes. Column 1: observing dates. Column 2: array configuration. Columns 3 and 4: the ranges of rest frequencies covered by the two spectral windows. Column 5: channel width. Columns 6–8: bandpass, time dependent gain, and absolute flux calibrators. Fig. 1. Left panel: full-color optical image of the Leo Triplet obtained with a 10.6-cm Takahashi FSQ106EDX apocromatic refractor (see Sect. 2.2). Right panel: peak intensity image of the Hispectra in the Leo Triplet system. The galaxies of NGC 3628, M 65, and M 66 are labeled in black and the Histructures beside the three galaxies are labeled in red. Here we also roughly estimate the flux recovered by the combined data. The integrated line fluxes are all obtained by integrating from VHEL =506.7 to 1125.1 km s−1(see Sect. 3). Then the total flux of the combined data is found to be 220.5Jy km s−1by spatially integrating the pixels with signal-to-noise ratios (S/Ns) larger than 5. Within the same pixels, the total flux of the VLA data is 185.6 Jy km s−1. The total flux of the Arecibo data within the pixels with S/Ns larger than 5 is 233.1 Jy km s−1. Therefore, the combined data have recovered most of the flux filtered by the VLA. The total Himass of the Leo Triplet system is 6.6×109Mby using the expression MHI =2.36 ×105D2 Mpc RSνdν(van Gorkom et al. 1986), which is consistent with the results reported by Stierwalt et al. (2009) (≈9×109M) and Haynes et al. (1979) (≈3×109M) for the distance of 11.3Mpc adopted here. In this paper, if not otherwise noted, Himaps represent the combined data. The velocity-integrated emission (zeroth moment), intensity-weighted velocity (first moment), and intensity-weighted velocity dispersion (second moment) maps are all constructed using the routines in the GILDAS software package3. A 3σch threshold, where σch is the rms noise 3https://www.iram.fr/IRAMFR/GILDAS level in a channel, is used to derive the first and second moment maps in order not to emphasize noisy features. 2.2. Optical observations An optical wide-field image of the Leo Triplet was obtained using high throughput clear filters with near-IR cut-off, known as luminance filters with a Takahashi FSQ106EDX 106 mm F/5 apochromatic refractor at native focal length (530mm) operated in Cuenca and Guadalajara (Spain) during 2019. It used an Atik 16200 CCD camera and Astrodon LRGB filters with a pixel scale of 200 .34. The reduction was carried out with the usual steps of bias and dark subtraction. The flat-fielding was done in Pixinsight, using a Geoptik 200-mm flat-field generator. Astrometry was obtained using SCAMP (Bertin 2006). A final stacked image (see Fig. 1, left panel) was obtained by combining the 25×600-s best images in luminance and the 18 ×300-s best images in RGB filters, with a total exposure time of 520min (8.40 h). 3. Results We first present the peak intensity (the intensity of the strongest channel) image of the Hidata to show the overall distribution A25, page 3 of 18 A&A 658, A25 (2022) Fig. 2. Hivelocity-integrated emission (zeroth moment, left panel) and intensity-weighted velocity (first moment, right panel) maps. The integration range is from VHEL =506.7 to 1125.1 km s−1to cover all the Hifeatures. In both panels the contour levels are set to (5,15,25,35,...,145) × 0.055 Jy beam−1km s−1. The galaxies NGC3628, M 65, and M 66 are labeled in both panels and a tiny filled ellipse in the lower left shows the beam. The red line in the upper left illustrates the 50kpc scale at a distance of 11.3 Mpc. Three blue dashed ellipses in the left panel demonstrate the main Hiemission regions of the three galaxies. of all the various Hicomponents and to highlight the narrow features in the right panel of Fig. 1. Here, we see that several structures, that is the plume, NGC 3628W, NGC 3628S, NGC 3628SW, IC 2767, M 66SE, and M 65S (named after their relative positions except the plume and IC2767), in addition to NGC 3628, M 65, and M 66, can clearly be identified and also labeled. We should note that the peak intensity map in the right panel of Fig. 1biases the Hiemission against that of galaxies with broad and irregular profiles. Therefore, the Hizeroth moment (integrated intensity) and first moment (velocity) maps are presented in Fig. 2to show the overall distributions of the Hiemission and velocity distributions in the system. In the left panel, contours and the color image show the zeroth moment map. The integration range is from VHEL =506.7 to 1125.1 km s−1to cover all the Hifeatures. The contour levels are set to (5,15,25,35,45,...,145) × σJy beam−1km s−1. The uncertainty of the integrated image σ is set to σch ×√Nch =0.055 Jy beam−1km s−1, where σch is the typical noise in a single channel (see Sect. 2), and Nch is the number of channels in the integrated velocity range (Nch =28 for Fig. 2). We also use three blue dashed ellipses in the left panel to cover the main Hiemission regions of the three galaxies. More importantly, the same ellipses are also added to Figs. 3–9 to indicate the relative locations of the features. In the right panel of Fig. 2, the color image shows the first moment map. The contours overlaid show the zeroth moment map in the same way as the left panel. We can see in Fig. 2, NGC 3628, M65, and M 66 are all detected in the zeroth moment map (left panel), and all show clear velocity gradients in the first moment map (right panel). However, the weaker features revealed by the right panel of Fig. 1only appear in a relatively narrow velocity range. Due to the large σ, proportional to √Nch, only the plume can be seen in the vicinity of NGC 3628 in Fig. 2. To further investigate the features along the velocity axis, we present the channel maps in Fig. 3, each one covering 20.7 km s−1. The contours in all the panels start at 3σch and go up in steps of 9σch, where σch is the typical noise of the integrated intensity in an individual channel (see Sect. 2). The additional structures with narrow line widths, M 66SE (640.7–702.5 km s−1), NGC 3628S (764.3– 867.4 km s−1), NGC 3628SW (826.2–888.0 km s−1), M 65S (743.7–764.3 km s−1), the plume (805.6–929.2km s−1), and IC 2767 (1032.3–1114.7 km s−1) are also labeled in Fig. 34. The remaining structures are consistent with previous Arecibo Hiobservations (e.g., Stierwalt et al. 2009) but the angular resolution of our observations is about four times better. To emphasize these structures and their velocity distribution, we present, in Figs. 4–9, their zeroth moment and first moment maps within their specific velocity ranges mentioned above. In these figures, the top two panels present the full view of the distributions of Hiemission and velocity, while the two lower panels show zoomed images of these structures. We note that these integrated ranges do not fully cover the Hiemission in the three galaxies, which means only a part of HI emission in each of the galaxies is shown in Figs. 4–8. The plume. Figure 4presents the zeroth moment (left two panels) and the first moment (right two panels) maps of the plume. In all the panels, the integrated velocity range is from VHEL =795.3 to 939.5 km s−1to highlight the emission from the plume. The uncertainty of the velocity-integrated intensity σ=σch ×√Nch =0.027 Jy beam−1km s−1, where σch ≈ 0.01 Jy beam−1km s−1(see Sect. 2) and Nch =7. From Fig. 4we can see that there are plenty of condensations along the plume and the strongest one is located at the 4There is another Histructure appearing in the west of NGC 3628 which is very likely caused by a sidelobe (see Appendix A). A25, page 4 of 18 G. Wu et al.: Himapping of the Leo Triplet Fig. 3. Heliocentric velocity channel maps of the Hiemission. The contours in all the panels start at 3σch and go up in steps of 9σch, where σch =0.01 Jy beam−1km s−1(see Sect. 2). The galaxies NGC3628, M 65, M 66 are labeled in each panel in black. M 66SE (640.7–702.5 km s−1), NGC 3628S (764.3–867.4 km s−1), NGC 3628SW (826.2–888.0 km s−1), M 65S (743.7–764.3 km s−1), the plume (805.6–929.2 km s−1), and IC 2767 (1032.3–1114.7 km s−1) are also labeled (in red) in their detected channels. The three blue dashed ellipses demonstrate the locations of the three galaxies belonging to the Leo Triplet. eastern tip, which was proposed to be a tidal dwarf galaxy by Nikiel-Wroczy´ nski et al. (2014). The Hispectra in the plume are extremely narrow and there is little velocity variation along the plume. Our Arecibo data have a better velocity resolution of about 1.9 km s−1after smoothing three contiguous velocity channels (see Appendix B). As can be seen in the right two panels of Fig. B.1, the narrowest spectra have FWHM line widths of about 20 km s−1, evident in the middle of the plume. These characteristics have also been presented by previous Hiobservations (e.g., Haynes et al. 1979). However, our observations spatially resolve A25, page 5 of 18 A&A 658, A25 (2022) Fig. 4. Hivelocity-integrated emission (zeroth moment) and intensity-weighted velocity (first moment) maps of the plume. Top two panels: in the left panel, contours and also the color image show the zeroth moment map. The integration range is from VHEL =795.3 to 939.5 km s−1. The contour levels are set to (3,6,9,12,15,18,21,30,40,50,60,...,180) ×0.027 Jy beam−1km s−1. The green ellipses demonstrate the clumps detected in the plume. Right panel: color image shows the first moment map. The contours overlaid are the same as the ones in the left panel. The galaxies NGC 3628, M 65, M 66, and NGC 3628E (the plume) are also labeled in both panels and a filled ellipse in the lower left shows the synthesized beam. The red line in the top left illustrates the 50 kpc scale at a distance of 11.3 Mpc. The three blue dashed ellipses demonstrate the locations of the three galaxies. Bottom two panels: same as the top two panels but zooming into the region that is illustrated by the green dashed rectangle in the top left panel. The black solid and dashed lines in the lower left panel show the loci and the averaged width of the radial profiles in Fig. 11. the four Hiclumps revealed by Stierwalt et al. (2009) into about 11 condensations (see Table 2and Fig. 4). The condensations are identified by the astrodendro package5(Rosolowsky et al. 2008) with an intensity threshold (min_value) of 5σand a size threshold (min_npix) of 10pixels. The minimum intensity difference (min_delta) to be considered an independent entity (leaf or branch in the package) is set to 1σas suggested by the package. The 11 condensations along the plume are reminiscent of the chains of tidal dwarfs that were seen in the Dentist’s Chair galaxy (Weilbacher et al. 2002), the Tadpole galaxy (Tran et al. 2003), and in simulations of tidal tails (Wetzstein et al. 2007; Noreña et al. 2019). Here we want to emphasize features that have not been noticed before. From the zeroth moment map of Fig. 4we can see that the Hiemission at the eastern tip and middle of the plume shows two branches separated by centrally located weaker Hiemission. Moreover, from the first moment map we can see that there are clearly two different velocities in the two branches, 5http://www.dendrograms.org/ which are also evident in the Arecibo data in Fig. B.1. From a combined view of the zeroth and first moment maps, we find that the northern branch belongs to a not completely continuous elongated structure with velocities of about 850km s−1, while the stronger southern branch shows velocities closer to 900km s−1. The northern side of the plume exhibits more diffuse gas than the southern one. There is no smooth transition between the two velocity components. Instead, the velocity field looks more like an overlap of two spatially and kinematically distinct gas filaments, which are likely caused by two well-separated arms in the plume drawn out from NGC3628. This is further discussed in a more comprehensive manner in Sect. 4. NGC 3628S. Figure 5presents the zeroth moment (left two panels) and the first moment (right two panels) maps of NGC 3628S. In all the panels the velocity ranges from VHEL =754.0 to 878.7 km s−1. The uncertainty of the velocityintegrated intensity σ=σch ×√Nch =0.025 Jy beam−1km s−1, where σch ≈0.01 Jy beam−1km s−1(see Sect. 2) and Nch =6. From the zeroth moment map we can see that NGC 3628S extrudes from NGC3628 with a uniform intensity of around A25, page 6 of 18 G. Wu et al.: Himapping of the Leo Triplet Fig. 5. Same as Fig. 4, but for NGC 3628S. The velocity range is from 754.0 to 878.7km s−1. The contour levels are set to (3,6,9,12,15,18,21,30,40,50,60,...,180) ×0.025 Jy beam−1km s−1. 3σtoward the south. The strongest emission of about 12σis found at the southern tip. Several pixels with intensities of about 3σfurther to the east (≈50) are beyond the scope of this work because of their limited size and intensities. The first moment map shows an overall velocity gradient in a roughly north-south direction along NGC 3628S, but with large dispersions. According to the categories in Toomre & Toomre (1972) and assuming that there was a tidal interaction between NGC3628 and M 66, this part should be the “bridge”. However, this bridge seems to be too prominent with respect to simulations of encounters of galaxies with similar total masses (e.g., Rots 1978). According to the model provided by Toomre & Toomre (1972), the material in the bridge might eventually fall back to the original galaxy, that is NGC 3628 in this case. Again, assuming that this is a bridge caused by tidal interaction, the velocity gradient along NGC 3628S might be caused by the gas falling back. NGC 3628SW. Figure 6presents the zeroth moment (left two panels) and the first moment (right two panels) maps of NGC3628SW. In all the panels the velocity ranges from VHEL =815.9 to 898.3 km s−1. The uncertainty of the velocityintegrated intensity σ=σch ×√Nch =0.021 Jy beam−1km s−1, where σch ≈0.01 Jy beam−1km s−1(see Sect. 2) and Nch =4. A25, page 7 of 18 A&A 658, A25 (2022) Fig. 6. Same as Fig. 4but for NGC 3628SW. The velocity range is from 815.9 to 898.3km s−1. The contour levels are set to (3,6,9,12,15,18,21,30,40,50,60,...,150) ×0.021 Jy beam−1km s−1. From the zeroth moment map of Fig. 6, we can see that NGC 3628SW consists of several pieces of small Hiclouds and a strongest condensation reaching a peak at about the 12σlevel. These cloudlets seem to form an arc-like structure first oriented westand then southward and seem to be connected to NGC 3628. The strong compact condensation at the end of the arc-like structure may show a northeast-southwest velocity gradient with blue velocities in the northeast. M66 SE. Figure 7presents the zeroth moment (left two panels) and the first moment (right two panels) maps of M66SE. The velocity range is from VHEL =630.3 to 712.8 km s−1in all the panels. The uncertainty of the velocity-integrated intensity σ=σch ×√Nch =0.023 Jy beam−1km s−1, where σch ≈ 0.01 Jy beam−1km s−1(see Sect. 2) and Nch =5. From the zeroth moment map of Fig. 7we can see that the main part of M66SE consists of a strong round clump and a faint tail extending to the south. On its western side, there are also a few separate pixels with emission of around 3σ, which are beyond the scope of this paper because of their limited sizes and intensities. From the first moment map we can see that the dense clump presents a south-north velocity gradient with highest velocities in the north. Meanwhile this velocity gradient is reversed compared to the velocity gradient of M66 (Fig. 2). The faint tail in the south of the clump shows velocities of about 640 km s−1. We further discuss this structure in Sect. 4. IC 2767. Figure 8presents the zeroth moment (left two panels) and the first moment (right two panels) maps of IC2767. The velocity range is from VHEL =1022.0 to 1125.1km s−1 in all the panels. The uncertainty of the velocity-integrated A25, page 8 of 18 G. Wu et al.: Himapping of the Leo Triplet Fig. 7. Same as Fig. 4but for M 66SE. The velocity range is from 630.3 to 712.8km s−1. The contour levels are set to (3,6,9,12,15,18,21,30,40,50,60,...,170) ×0.023 Jy beam−1km s−1. intensity σ=σch×√Nch =0.023 Jy beam−1km s−1, where σch ≈ 0.01 Jy beam−1km s−1(see Sect. 2) and Nch =5. This Histructure is a compact source with an ≈34 km s−1arcmin−1east-west velocity gradient with blue velocities in the east. According to its location and velocity gradient, it is probably associated with the galaxy IC2767 as is also suggested by Stierwalt et al. (2009). The systemic velocity of IC2767 is 1080 km s−1(Haynes et al. 2011;Gavazzi et al. 2012) which is consistent with that of the Hiobservations. The distance of IC 2767 is much larger than that of the Leo Triplet. Gavazzi et al. (2012) and Haynes et al. (2011) suggested 19.6 and 24 Mpc, respectively, which means IC 2767 is not likely a member of the Leo Triplet. M 65S. Figure 9presents the zeroth moment (left two panels) and the first moment (right two panels) maps of M65S. The velocity range is from VHEL =733.4 to 774.6km s−1in all the panels. The uncertainty of the velocity-integrated intensity A25, page 9 of 18 A&A 658, A25 (2022) Observatory is a facility of the National Science Foundation operated under cooperative agreement by Associated Universities, Inc. The Arecibo Observatory is a facility of the National Science Foundation operated under cooperative agreement by the University of Central Florida and in alliance with Universidad Ana G. Mendez, and Yang Enterprises, Inc. References Anand, G. S., Lee, J. C., Van Dyk, S. D., et al. 2021, MNRAS, 501, 3621 Cole, S., Lacey, C. G., Baugh, C. M., et al. 2000, MNRAS, 319, 168 Martínez-Delgado, D., Gabany, R. J., Crawford, K., et al. 2010, AJ, 140, 962 Martínez-Delgado, D., D’Onghia, E., Chonis, T. S., et al. 2015, AJ, 150, 116 Afanasiev, V. L., & Sil’chenko, O. K. 2005, A&A, 429, 825 Arp, H. 1966, ApJS, 14, 1 Baan, W. A., & Goss, W. M. 1992, ApJ, 385, 188 Bekki, K., & Tsujimoto, T. 2019, ApJ, 886, 121 Bertin, E. 2006, ASP Conf. Ser., 351, 112 Chemin, L., Cayatte, V., Balkowski, C., et al. 2003, A&A, 405, 89 Chromey, F. R., Elmegreen, D. M., Mandell, A., et al. 1998, AJ, 115, 2331 Duan, Z. 2006, AJ, 132, 1581 Dumke, M., Krause, M., Beck, R., et al. 2011, ASP Conf. Ser., 446, 111 Gavazzi, G., Fumagalli, M., Galardo, V., et al. 2012, A&A, 545, A16 Haynes, M. P., Giovanelli, R., & Roberts, M. S. 1979, ApJ, 229, 83 Haynes, M. P., Giovanelli, R., Martin, A. M., et al. 2011, AJ, 142, 170 Ibata, R., Irwin, M., Lewis, G., et al. 2001, Nature, 412, 49 Jacobs, B. A., Rizzi, L., Tully, R. B., et al. 2009, AJ, 138, 332 Jennings, Z. G., Romanowsky, A. J., Brodie, J. P., et al. 2015, ApJ, 812, L10 Kormendy, J., & Bahcall, J. N. 1974, AJ, 79, 671 Kroupa, P. 2015, Can. J. Phys., 93, 169 Toomre, A., & Toomre, J. 1972, ApJ, 178, 623 Majewski, S. R., Skrutskie, M. F., Weinberg, M. D., et al. 2003, ApJ, 599, 1082 McMullin, J. P., Waters, B., Schiebel, D., et al. 2007, ASP Conf. Ser., 376, 127 Noreña, D. A., Muñoz-Cuartas, J. C., Quiroga, L. F., et al. 2019, Rev. Mex. Astron. Astrofis., 55, 273 Nikiel-Wroczy´ nski, B., Soida, M., Bomans, D. J., et al. 2014, ApJ, 786, 144 Pawlowski, M. S., Kroupa, P., & de Boer, K. S. 2011, A&A, 532, A118 Reid, M. J., & Honma, M. 2014, ARA&A, 52, 339 Renaud, F., Famaey, B., & Kroupa, P. 2016, MNRAS, 463, 3637 Reuter, H.-P., Sievers, A. W., Pohl, M., et al. 1996, A&A, 306, 721 Rosolowsky, E. W., Pineda, J. E., Kauffmann, J., et al. 2008, ApJ, 679, 1338 Rots, A. H. 1978, AJ, 83, 219 Sault, R. J., Teuben, P. J., & Wright, M. C. H. 1995, ASP Conf. Ser., 77, 433 Soida, M., Urbanik, M., Beck, R., et al. 2001, A&A, 378, 40 Stierwalt, S., Haynes, M. P., Giovanelli, R., et al. 2009, AJ, 138, 338 Tran, H. D., Sirianni, M., Ford, H. C., et al. 2003, ApJ, 585, 750 van Gorkom, J. H., Knapp, G. R., Raimond, E., et al. 1986, AJ, 91, 791 Weilbacher, P. M., Fritze-v. Alvensleben, U., Duc, P. A., et al. 2002, ApJ, 579, L79 Wetzstein, M., Naab, T., & Burkert, A. 2007, MNRAS, 375, 805 We˙zgowiec, M., Soida, M., & Bomans, D. J. 2012, A&A, 544, A113 Wilding, T., Alexander, P., & Green, D. A. 1993, MNRAS, 263, 1075 Young, J. S., Tacconi, L. J., & Scoville, N. Z. 1983, ApJ, 269, 136 Zhang, X., Wright, M., & Alexander, P. 1993, ApJ, 418, 100 Zwicky, F. 1956, Ergebnisse der exakten Naturwissenschaften, 29, 344 A25, page 16 of 18 G. Wu et al.: Himapping of the Leo Triplet Appendix A: NGC 3628W NGC 3628W: Fig. A.1 presents the first moment (left panel) and the second moment (right panel) maps of NGC 3628W. The velocity range is from VHEL =609.8 to 1063.2 km s−1in both panels. The uncertainty of the velocityintegrated intensity σ=σch ×√Nch =0.048 Jy beam−1km s−1, where σch ≈0.01 Jy beam−1km s−1(see Sect. 2) and Nch =22. As we can see in Fig. A.1, the morphology and velocity patterns of this structure are very similar to the strong central part of NGC3628, that is they show a similar elongated morphology and velocity gradient. This feature is only present in the VLA data but is absent in the Arecibo data. Thus it is likely caused by the VLA sidelobes. It was also not seen in any other study in spite of its strength. Thus, we conclude that NGC 3628W is very likely a spurious feature. As a consequence, it is not part of our analysis. Fig. A.1. Same as Fig. 4but for NGC 3628W. The velocity range is from 609.8 to 1063.2 km s−1in both panels. The contour levels are set to (3,6,9,12,15,18,21,30,40,50,...,160) ×0.048 Jy beam−1km s−1. A25, page 17 of 18 A&A 658, A25 (2022) Appendix B: The Arecibo Observations We first masked the strong spikes in the spectra caused by radio frequency interference (RFI) and smoothed three contiguous velocity channels to enhance the S/N of the individual channels in the Arecibo data. The velocity resolution and the noise level of the smoothed data turn into 1.9km s−1and 2.0 mJy beam−1, respectively. Our noise level is slightly better than that in Stierwalt et al. (2009) (≈4 mJy beam−1scaled to 1.9 km s−1). Figure B.1 shows the zeroth moment (left panels), first moment (middle panels), and second moment (right panels) of the Leo Triplet (top three panels) and the plume (lower three panels). The velocity ranges of the top and lower panels are from VHEL =506.7 to 1125.1 km s−1and 795.3 to 939.5 km s−1, respectively, which are the same as those for Fig. 2 and 4. The uncertainties of the upper and lower zeroth moment maps are 0.07 Jy beam−1km s−1and 0.03 Jy beam−1km s−1. A threshold of 3σch, where σch is the noise level in a channel, is used to derive the first and second moment maps to avoid emphasizing contributions from spectral noise and the RFI. We can see in the top left panel of Fig. B.1 that our Arecibo map is consistent with that presented by Stierwalt et al. (2009). We note that an about 100-long spur, reported by Stierwalt et al. (2009) in the north of NGC 3628 extending further north, is not evident in our map. We checked our data and found a very narrow andfaintfeaturewith velocitiesofabout950 km s−1,whichissimilar to that reported by Stierwalt et al. (2009). But this feature is not discussed in our paper due to its poor S/N. Meanwhile, M 65S has also been detected by Stierwalt et al. (2009), but this feature extends further north in their map. These discrepancies might be partially caused by larger noise levels at the edges of our Arecibo images. From the middle panels, we can see that the Arecibo data show quite a similar velocity pattern as that seen in Figs. 2and 4. In the plume, there are also two velocity regimes with similar spatial distributions as in Fig. 4. Since the Arecibo data have a better velocityresolution thanthecombined data,wefurther providethe second moment maps in Fig. B.1. From the two maps, we can see that the narrowest spectra are located in the middle of the plume, showing linewidths of about 20 km s−1, which is consistent with the observations in Haynes et al. (1979) and the velocity resolution of our combined Arecibo and VLA data. Fig. B.1. HI data exclusively from the Arecibo telescope with 30 .3 angular resolution. Top three panels: The Hivelocity-integrated emission (zeroth moment), intensity-weighted velocity (first moment), and intensity-weighted velocity dispersion (second moment) maps of the Leo Triplet. In the left panel, contours and also the color image show the zeroth moment map. The integration range is from VHEL =506.7 to 1125.1 km s−1. The contour levels are set to (3, 6, 9, 12, 18, 24, 30, 36, 42, 48, 54, 100, 150, 200, ..., 600) ×0.07 Jy beam−1km s−1. In the middle and right panels, the color image shows the first and second moment maps. The contours overlaid are the same as the ones in the left panel. The galaxies NGC3628, M 65, and M 66 are also labeled in both panels and a filled ellipse in the lower left shows the synthesized beam. Bottom three panels: Same as the top panels, but exclusively for the plume. The integration range is from VHEL =795.3 to 939.5 km s−1. The contour levels are set to (3, 15, 27, 39, 51, 63, 75, 87, 99, 111, 133, 200, 250, 300, ..., 600) ×0.03 Jy beam−1km s−1. The red line in each panel illustrates the 50kpc scale at a distance of 11.3 Mpc. A25, page 18 of 18