arXiv:1610.00495v1 [astro-ph.GA] 3 Oct 2016 cetaglrrslto fteH the of suf- resolution and medium angular data interstellar ficient supporting multi-wavelength surrounding With the recently-formed (ISM). and between interplay the massive within of Systems study detailed internal halos. the low- of ∼ govern understanding that our processes probing for oratories set fglx vlto,fo asv trformation massive the revolu- from diverse evolution, have ; and of surveys multiple low-mass aspects on these of touches approach whole, understanding panchromatic al. a et our Kennicutt as infrared 2009; tionized Taken the al. et in the Dale 2011). 2003; surveys and al. (2009), KINGFISH et al. (Kennicutt opti- and et HST LVL, Dalcanton the by (2015), SINGS, al. ANGST et survey LE- Calzetti (HST) cal by Telescope survey Space by Hubble in- UV survey the GUS (2011), imaging list al. (UV) incomplete et ultraviolet Lee An span GALEX that the surveys cludes spectrum. major electromagnetic recent the in attention con- matter. significant and baryonic sources to the dark of of dynamics ratios the the strain probe also can we a 0Mccnb eovdit niiulsas allowing stars, individual into resolved be can Mpc 10 etb soitdUieste,Inc. th ag Universities, of cooperative Associated facility under by a ment operated is Foundation Observatory Science USA Astronomy National 47907, Radio IN National Lafayette, The West Avenue, Northwestern 525 USA 55104, MN Paul, Saint Avenue, Hewitt 02143 MA Somerville, Avenue, USA USA 22903-2475, VA 55105, MN Paul, Saint Avenue, Grand 1600 ery trfrigdafglxe r motn lab- important are galaxies dwarf star-forming Nearby, h wr aaisi h oa ouehv received have volume local the in galaxies dwarf The cetdfrpbiaini h srnmclJournal L Astronomical using the typeset Preprint in publication for Accepted onM Cannon M. John Brock 5 4 3 2 1 OAINLDNMC N TRFRAINI H ERYDAFG DWARF NEARBY THE IN FORMATION STAR AND DYNAMICS ROTATIONAL eateto hsc n srnm,Pru University, Purdue Astronomy, and Physics of Department 1536 University, Hamline Highland MS-B1807, Physics, 81 of Department School, High Somerville Department, Science Charlottesville Road, College, Edgemont 520 Macalester Charlottesville, NRAO Astronomy, & Physics of Department ul-ihrrlto n otxulz h ytmaogtohrlo other amongst system headings: the Subject contextualize and relation Tully-Fisher eoiyaayi,w ne oainlvlct f31 of velocity rotational a infer we analysis, velocity a-VadH and far-UV uniytedge fc-ptaiyo es H dense of co-spatiality of degree the quantify ihH high h alG asyVr ag ra (VLA Array Large Very Jansky G. Karl the ulte fisnurlitrtla eimhv eandunstudied. remained have medium unt interstellar but neutral surveys, galaxy its nearby of recent and qualities ongoing many in included aeu n tla opnns h H The components. stellar of and scales gaseous physical on disk etro h tla ouain.TeH The populations. stellar the of center sa cieysa-omn aaywt iepedH widespread with galaxy star-forming actively an is 1 ra .Eisner A. Brian , epeetnwH new present We I assraedniydslyacecn-hpdmrhlg that morphology crescent-shaped a display density surface mass 1. 1 nrwT McNichols T. Andrew , A INTRODUCTION T α E tl mltajv 5/2/11 v. emulateapj style X msin h eta a ieaisaecmlx sn spatially-r a using complex; are kinematics gas neutral The emission. aais vlto aais wr aais reua aais in galaxies: — irregular galaxies: — dwarf galaxies: — evolution galaxies: NC5238) (NGC 1 ahenFitzgibbon Kathleen , I pcrlln mgso h erylwms aayNC53,acquir 5238, NGC galaxy low-mass nearby the of images line spectral I ∼ cetdfrpbiaini h srnmclJournal Astronomical the in publication for Accepted 1c pcrlline, spectral 21-cm 0 c loigu oudraeadtie oprtv td fthe of study comparative detailed a undertake to us allowing pc, 400 Ruvolo 1,2 ao .Teich G. Yaron , I I oundniyeces10 exceeds density column iki smerci h ue ein,adteaesof areas the and regions, outer the in asymmetric is disk 1 1 a uno Singer Quinton , ree- aa Miazzo Masao , .Lctda itneo 4.51 of distance a at Located ). Abstract e I , a n ie fogigsa omto stae by traced as formation star ongoing of sites and gas ± eta yrgnpoete ftesa-omn dwarfs star-forming universe. the the local of the of in census properties observational hydrogen complete neutral nearly a provided aaySre;Hne ta.21) n VI teLo- (the LVHIS and H 2012), Volume al. cal et Hunter Survey; Galaxy k uvy(DS mgs hc eelabu stel- blue clusters; a stellar reveal prominent numerous which and images, population lar (SDSS) Survey the Sky of observations detailed for ISM. target neutral make interesting formation an star active it and work. mass, present low neutral the proximity, until The published its been not of have hydrogen Re- measurements manuscripts. interferometric peer-reviewed ap- 100 markably, than has more galaxy Dreyer in star-forming in peared actively cataloged nearby, 1479, Originally this Mrk (1888), 828, 64). VV is Zw afore- I 8565, system UGC the 1332+518, such as SBS from known One (e.g., (also 5238 infrared surveys). NGC the galaxy to nearby observa- UV mentioned stud- multi-wavelength the H from extensive been tions have the yet also in not systems detail have in that ied galaxies volume local irregular multiple exist dwarf there criteria, selection ferent ofebc ordainblnei h ISM. the in balance radiation to feedback to e;Ote l 02,LTL-HNS(oa Irregu- H (Local The in LITTLE-THINGS Extremes Luminosity 2012), Trace Galax- That al. Treasury et lars Survey Ott Galaxy Nearby ies; Array ACS 2016, Large of (Very al. et VLA-ANGST Survey Teich 2016), H 2011, al. et of al. McNichols et Survey Cannon (the Dwarfs; SHIELD Low-mass 2008), 2002), Begum Survey; Swaters al. GMRT et Galaxies; Galaxies Irregular Spiral (Faint multiple and FIGGS in Irregular H (Westerbork of detail WHISP Survey significant surveys: in interferometric studied major been has lation α ms km 5 ic aho h frmnindH aforementioned the of each Since G 28i eovdit tr nSonDigitized Sloan in stars into resolved is 5238 NGC rmteH the From n Vcniumeiso.Tesuc is source The emission. continuum UV and 1,3 ahrn Ball Catherine , − 1 saNizami Asra , 1 1 epaeNC53 ntebaryonic the on 5238 NGC place We . I lti oktesailyresolved spatially the work this il -asgalaxies. w-mass 21 uvy ib 02.Teesreshave surveys These 2012). Kirby Survey; I esetv,tenab aaypopu- galaxy nearby the perspective, cm u H Our − 2 ssihl ffe rmthe from offset slightly is 1 1 rde Reilly Bridget , nmc fteds.We disk. the of much in I onBanovetz John , ± pcrlln.Sm fthese of Some line. spectral .4Mc G 5238 NGC Mpc, 0.04 I mgsrslethe resolve images sle position- esolved LX G 5238 NGC ALAXY I dividual dwith ed uvy a dif- has surveys I 1 1,4,5 Elizabeth , nExtremely in Annika , I Nearby I 2

2.1. HI Observations Table 1 Basic Characteristics of NGC 5238 HI spectral-line observations of NGC5238 were ac- quired with the Karl G. Jansky Very Large Array (VLA) Parameter Value in the C configuration on February 1, 2016 for program Right ascension (J2000) 13h 34m 42.s5 TDEM0021. This program was executed under the aus- Declination (J2000) +51◦36′49′′ pices of the “Observing for University Classes” program, Adopted distance (Mpc) 4.51 ± 0.04a a service provided by the NRAO as an opportunity for b MB (Mag.) −14.61 7c courses teaching radio astronomy theory to acquire new M⋆ (M⊙) 8.9 × 10 −1 d observations to be analyzed by students. All of the re- Single-dish SHI (Jy km s ) 5.77 ± 0.61 −1 e sults in this manuscript stem from the efforts of under- Interferometric SHI (Jy km s ) 4.56 ± 0.46 7d graduate students at Macalester College. Single-dish H I mass MHI (M⊙) (2.8± 0.3) × 10 7e Interferometric H I mass MHI (M⊙) (2.2± 0.3) × 10 The WIDAR correlator divides a 4.0 MHz bandwidth into 1,024 channels, delivering a native spectral resolu- −1 −1 a Tully et al. (2009) tion of 0.86 km s ch . The primary and phase cal- b Modifying MB from Cook et al. (2014a) for the adopted ibrators were J1331+305 (3C286) and J1400+621, re- distance. spectively. The total on-source integration time was ap- c Modifying M⋆ from Cook et al. (2014b) for the adopted proximately 1.6 hours. The VLA data were calibrated distance. 6 7 d using standard prescriptions in the AIPS and CASA Thuan et al. (1999); note that those authors apply a +16% increase in the observed flux integral (4.98 ± 0.53 environments. Imaging of the J1400+6210 phase cali- Jy km s−1) for beam effects. brator field yielded a flux density Sν = 4.24 ± 0.01 Jy. e This work. Continuum subtraction of the NGC 5238 field was per- the central cluster shows very strong nebular emission formed in the uv plane using a first-order fit to line-free lines. Broad-band HST images are used to derive a dis- channels bracketing the galaxy in the central 50% of the tance of 4.51 ± 0.04 Mpc in Tully et al. (2009); we adopt bandpass. this distance throughout the present work, and show Imaging of the calibrated uv visibilties followed the resulting HST images below. NGC 5238 is included standard prescriptions similar to those described in in the aforementioned LVL (Dale et al. 2009), GALEX Cannon et al. (2015) and in Teich et al. (2016). In brief, UV (Lee et al. 2011), and LEGUS (Calzetti et al. 2015) the input continuum-subtracted database was spectrally smoothed by a factor of two, to produce a velocity res- surveys. Moustakas & Kennicutt (2006) use strong- − − − line methods to derive a gas-phase oxygen abundance olution of 7.812 kHz ch 1 (1.56 km s 1 ch 1). The cube 12 + log(O/H) = 7.96 ±0.20 (see also Marble et al. 2010), was then inverted and cleaned using the IMAGR task corresponding to Z ≃ 19%Z⊙ using the Solar oxygen in AIPS; the ROBUST factor is 0.5. Cleaning was per- abundance of Asplund et al. (2009). Modifying the rele- formed to 2.5 times the rms noise per channel in line- vant values derived in those works for the updated dis- free channels. Residual flux rescaling was enforced dur- ing the imaging process (Jorsater & van Moorsel 1995). tance, we collect global physical properties of NGC 5238 ′′ ′′ in Table 1. The resulting synthesized beam size of 17.69 × 15.40 was smoothed to a circular 18′′ beam. The rms noise in The neutral hydrogen properties of NGC5238 have ′′ − only been studied with single-dish observations. Using the final 18 cube is 1.4 mJy Bm 1. the Nan¸cay 300-m telescope, Thuan et al. (1999) find Two-dimensional moment maps were produced from −1 the three-dimensional datacube as follows. The original W50 = 32 ± 4 kms and an observed H I line inte- ′′ ′′ −1 17.69 × 15.40 cube was first spatially smoothed by a gral SHI = 4.98 ± 0.53 Jy km s (corrected to 5.77 ± 0.61 ′′ Jykms−1 for beam effects). At our adopted distance, Gaussian kernel to a beam size of 30 . Using the rms NGC 5238 is a relatively low H I mass system, with noise in this cube, a threshold mask was applied at the 7 2.5 σ level. The resulting cube was then examined by MHI = (2.8±0.3) × 10 M⊙; including a 35% correc- tion for Helium and other metals, the total gas mass of hand to isolate real emission, which is required to be both 7 spectrally and spatially coincident across three neighbor- NGC 5238 is (3.7±0.3) × 10 M⊙. The source is compa- ing channels. The resulting blanked cube was used as a rable to the most massive galaxies in the SHIELD sam- ′′ ple (Teich et al. 2016; McNichols et al. 2016); that pro- transfer mask against the 18 beam cube. The result was gram explicitly studies systems that inhabit the impor- collapsed to create traditional moment maps represent- ing H I mass surface density, intensity weighted velocity tant but previously understudied mass range log(MHI) <∼ . field, and velocity dispersion. 107 2, and we refer the reader to Teich et al. (2016) and McNichols et al. (2016) for details. 2.2. Multiwavelength Archival Observations We organize this paper as follows. § 2 presents de- We compare our new H I images with archival ground tails about the new H I observations and the supporting and space-based imaging in order to study the stellar datasets. § 3 presents the H I properties of NGC 5238, populations in NGC 5238. Calibrated images in the including a quantitative comparison of the degree of co- SDSS g, r, and i filters were obtained from the SDSS spatiality between rate tracers (Hα and public website (York et al. 2000). Archival HST images far-UV emission) and H I mass surface densities. § 4 ex- from programs 10905 (Advanced Camera for Surveys, amines the complex neutral gas kinematics of NGC 5238. F606W and F814W filters) and 13364 (Wide Field Plane- In § 5 we contextualize NGC 5238 against the SHIELD tary Camera 3, F275W, F336W, F438W filters, from the galaxies, and in § 6 we draw our conclusions. 6 Developed and maintained by NRAO 2. OBSERVATIONS AND DATA HANDLING 7 https://casa.nrao.edu 3

(a) (b)

Figure 1. SDSS 3-color image of NGC 5238; SDSS g,r,i filters are blue, green, and red, respectively. NGC 5238 is marginally resolved into individual stars in ground-based images. LEGUS survey; Calzetti et al. 2015) allow us to study In Figure 1 we present a first-order comparison of the the stellar populations at the highest angular resolutions, gaseous and stellar components of NGC5238. The H I from the near-UV to the near-IR. We use these images mass surface density distribution (discussed in more de- to construct color mosaic images of NGC5238. The LE- tail below) is overlaid as contours on a 3-color SDSS im- GUS data (Calzetti et al. 2015) highlight the young stel- age. The outer H I disk is asymmetric; the morphologi- lar populations, while the optical data highlight the older cal major axis is elongated from northeast to southwest. stellar populations (and some nebular emission). The high column density H I gas shows a crescent-shaped Continuum-subtracted Hα images are acquired from morphology. The H I surface density maximum is not co- the LVL survey data products (Kennicutt et al. 2008). incident with the central optical peak, but rather is offset The Bok 2.3m telescope was used to image NGC5238 in to the northeast by ∼300 pc; such offsets are not unusual the standard Hα and R-band filters. A bright foreground in dwarf galaxies (see, e.g., van Zee et al. 1997; Hunter star is located in close angular proximity to the central etal. 1998). The H I disk is slighlty larger than the high stellar cluster in NGC 5238. This object leaves artifacts surface brightness optical body; the SDSS Petrosian-r ra- from continuum subtraction in the field; we discuss this dius (45′′) can be compared to the size of the H I disk source in more detail below. The total Hα luminosity of (∼90′′ radius at the level of 1020 cm−2). NGC5238 derived by Kennicutt et al. (2008) is corrected The H I spectral line is detected at high significance for foreground . Scaling this value from 5.2 across roughly 25 channels (∼40 kms−1) of the final 18′′ Mpc to the new distance of 4.51 Mpc, we find log(LHα) datacube. Figure 2 shows individual channel maps of −1 = 39.09, corresponding to log(SFRHα)= −2.01 M⊙ yr the H I emission across which gas is detected. Summing using the star formation rate calibration from Kennicutt the emission across these channels produces the global (1998). H I profile shown in Figure 3. As expected based on GALEX far-UV (FUV) and near-UV (NUV) imaging the galaxy’s low mass, the profile is Gaussian-like with of NGC5238 is presented in Lee et al. (2011). The source no obvious double-horn structure. Using the intensity- is detected at high significance in both filters, indicat- weighted average of the global H I profile, the systemic −1 ing recent star formation over the last few hundred Myr. velocity is VHI = 232 ± 1 kms ; this is in good agree- −1 Lee et al. (2011) calculates the total extinction-corrected ment with VHI 235 ± 2 kms derived in Thuan et al. FUV magnitude from NGC 5238, mFUV = 15.19 ± 0.05. (1999). Using the star formation rate calibration derived in Integrating the detected flux from NGC 5238 produces McQuinn et al. (2015), this corresponds to log(SFRFUV) the moment maps shown in Figure 4. As noted above, −1 21 = −1.81 M⊙ yr at our adopted distance; note that this the high mass surface density H I gas (NHI >∼ 10 star formation rate is slightly higher than using the star cm−2, shown by the white contour in Figure 4) displays formation rate metric from Hao et al. (2011), which gives a crescent-shaped morphology. The peak H I column −1 21 −2 log(SFRFUV) = −2.00 M⊙ yr . density is 1.67 × 10 cm , which corresponds to an H I −2 mass surface density of 13.4 M⊙ pc . We discuss the 3. NEUTRAL GAS AND STAR FORMATION IN NGC 5238 H I mass surface density and its relation to the ongoing § 3.1. The HI Properties of NGC 5238 and recent star formation in more detail in 3.2. The H I moment one map, representing intensity- 4

51 39 252.8 km/s 251.1 km/s 249.5 km/s 247.8 km/s 246.1 km/s 38 37 36 51 39 244.5 km/s 242.8 km/s 241.2 km/s 239.5 km/s 237.9 km/s 38 37 36 51 39 236.2 km/s 234.6 km/s 232.9 km/s 231.3 km/s 229.6 km/s 38 37 36 51 39 Declination (J2000) 228.0 km/s 226.3 km/s 224.7 km/s 223.0 km/s 221.4 km/s 38 37 36 51 39 219.7 km/s 218.1 km/s 216.4 km/s 214.8 km/s 213.1 km/s 38 37 36

13 34 50 40 35 30 13 34 50 40 35 30 13 34 50 40 35 30 Right Ascension (J2000)

Figure 2. Channel maps of NGC 5238. Contours are overlaid at 3σ, 6σ, and 12σ, where σ is the rms noise in line-free channels of the HI cube (σ = 1.4 × 10−3 JyBm−1). The circular 18′′ beam size is shown in the bottom left of the first panel. weighted H I velocity, is shown as panel (b) of Figure 4. sions ranging from 7-12 kms−1, in good agreement with As discussed in detail in McNichols et al. (2016), the col- characteristic values determined from H I observations lapse of the three-dimensional cube to two dimensions of many types of galaxies (e.g., Tamburro et al. 2009; results in a more narrow velocity width of the source; by Ianjamasimanana et al. 2015; Mogotsi et al. 2016) and in comparing Figures 2 and 4, it is clear that H I gas is particular for values determined for local volume dwarf detected over ∼40 kms−1 in the data cube but that the galaxies (e.g., Hunter et al. 2012; Ott et al. 2012). We resulting velocity field only captures ∼20 kms−1. Ex- note that the H I column density maximum is coinci- aming the velocity field, the H I kinematics are complex. dent with gas with σ ≃ 8 kms−1. Similarly, in the inner There is a velocity gradient extending from southwest regions of the disk with active star formation, the H I to northeast along a position angle of roughly 20◦ (east velocity dispersion remains below 10 km s−1. of north). However, significant velocity asymmetries are In terms of H I content, NGC5238 can be directly present throughout the disk; there is an S-shaped veloc- compared with the sample of 12 SHIELD galaxies pre- ity profile that may be suggestive of a warp in the disk. sented in Teich et al. (2016) and McNichols et al. (2016). We discuss the neutral gas kinematics in detail in § 4. NGC5238 has an H I mass comparable to the most mas- The H I moment two map shows velocity disper- sive SHIELD galaxies; only AGC 749237 is more H I- 5

Figure 3. Global H I profile of NGC 5238, created by summing the flux in each channel of the 18′′ resolution blanked datacube. The H I −1 systemic velocity, derived from an intensity weighted mean of the spectrum, is VHI 232 ± 1 kms . massive. The peak H I column density in NGC 5238 shown in panel (a) clearly highlights the youngest stel- is equal to the peaks found in AGC110482 and in lar clusters. The most prominent of these structures is AGC 749237. In the following sections we further con- located in the southern region of the galaxy (coincident textualize NGC5238 against the SHIELD sources. with the southern-most extension of the 1.4 × 1021 cm−2 3.2. Comparing Stellar Population and Neutral Gas column density contour); nebular emission (arising from Properties the Hα spectral line, which falls in the F606W filter) sur- rounds this cluster in the optical image shown in panel NGC 5238 is an actively star-forming (see (b). Smaller young clusters, identified by a comparison § 2.2). Using Hα emission as a recent star formation rate of panels (a) and (b), are located in the regions of high- indicator (within the last 10 Myr), log(SFRHα)= −2.01 est H I column density (N > 16 × 1020 cm−2, in the −1 HI ∼ M⊙ yr . Averaged over a longer timescale of ∼100 northeast region of the optical body) as well as in the op- Myr, the GALEX far-UV emission indicates a star for- tical center. A close inspection of panel (b) reveals that mation rate that is ∼60% higher: log(SFRFUV)= −1.81 nebular emission is associated with all of these clusters, −1 M⊙ yr . Each of these star formation rates are higher and is brightest near the central, bright cluster. than those of any of the SHIELD galaxies. As expected, the nebular emission visible in the To compare the neutral gas properties with the loca- HST image matches the morphology of the continuum- tions of recent star formation in NGC 5238, in Figure 5 subtracted Hα image shown in panel (d). The ground- we show multi-wavelength imaging compared to the H I based Hα image has better surface brightness sensitivity mass surface density. The fields of view are the same due to the comparatively narrow filter width; this image in all panels. The H I column density contours high- clearly reveals widespread Hα emission throughout the − light regions above 1021 cm 2. This mass surface den- galaxy. The most Hα-luminous region is the central stel- sity level has been empirically identified as the requisite lar cluster, followed closely by the strong Hα emission level for massive star formation as traced by Hα emis- from the young stellar cluster in the southern disk. In sion (Skillman 1987); note that more recent studies have addition to these two major maxima, there is Hα emis- shown that such a surface density is not universal (e.g., sion throughout the disk (including regions associated Bigiel et al. 2008, Elmegreen & Hunter 2015, Teich et al. with obvious UV clusters in the LEGUS HST image) 2016, and the various references therein). The aforemen- and also multiple dramatic loops and arcs that extend tioned crescent-shape morphology is very prominent in significantly beyond the locations of the most luminous the highest column density gas. clusters (for example, the nearly circular structure ex- The HST images shown in panels (a) and (b) of Fig- tending to the western edge of the disk). ure 5 dramatically resolve the stellar populations in NGC 5238 harbors widespread UV emission in both NGC5238. The LEGUS image (Calzetti et al. 2015) 6

HI Column Density (cm −2 ) Velocity (km s −1 ) Velocity Dispersion (km s −1 ) 0 5 10 15 225 230 235 240 8 10 12

51 38 00 51 38 00 51 38 00

37 30 37 30 37 30

00 00 00

36 30 36 30 36 30 Declination (J2000) Declination (J2000) Declination (J2000)

00 00 00 (a) (b) (c) 35 30 35 30 35 30 13 34 50 48 46 44 42 40 38 13 34 50 48 46 44 42 40 38 13 34 50 48 46 44 42 40 38 36 34 Right ascension (J2000) Right ascension (J2000) Right ascension (J2000)

Figure 4. H I moment maps of NGC 5238. Panel (a) shows the H I mass surface density in units of atoms cm−2; contours are at levels of (1,2,4,8,16) × 1020 cm−2. Panel (b) shows the intensity-weighted velocity field; contours span the range 224 – 238 km s−1, spaced by 2 km s−1 per contour. Panel (c) shows the velocity dispersion of the H I gas; contours span the range 7 – 11 km s−1, spaced by 1 km s−1 per contour (note that this separation is formally smaller than our spectral resolution). The 18′′ circular beam size is shown in the bottom left of each panel. the GALEX FUV and the NUV bands. As panel (c) of molecular ISM (e.g., Skillman 1987; Bigiel et al. 2008; Figure 5 shows, the UV emission peaks in two prominent Roychowdhury et al. 2014; Elmegreen & Hunter 2015; clusters that match with the highest surface brightness Teich et al. 2016); no single prescription has yet been Hα regions and with the the prominent UV clusters in the identified that holds in all environments. For uniformity, LEGUS images. There is also widespread UV emission thus we follow the procedures outlined in Teich et al. throughout the rest of the stellar disk, including emission (2016) by deriving the index of the “Schmidt-Kennicutt” from young, blue stars that trace the loop structure that relation (Schmidt 1959), which relates a star formation is seen in Hα on the western side of the disk. rate surface density to a gas surface density: When interpreting Figure 5, it is important to note N that there is a bright foreground star in close angular ΣSFR ∝ (Σgas) (1) proximity to the central young cluster. This object ap- pears somewhat pink in the LEGUS image in Figure 5, where the star formation rate surface density (ΣSFR) is −1 −2 and is evident as an imperfectly-subtracted image arti- in units of M⊙ yr kpc , the gas surface density (Σgas) −2 fact in the Hα image. The SDSS spectroscopic aperture in units of M⊙ pc , and N is a positive number. Specif- contains this source; emission lines are nonetheless very ically, we compare the star formation rate surface densi- prominent in the resulting spectrum (not shown here), ties (hereafter, ΣFUV SFR for FUV and ΣHα SFR for Hα) which is centered on the Hα surface brightness maximum against the H I mass surface densities (hereafter, ΣHI) that is slightly offset to the east. on both a radially-averaged and on a pixel-by-pixel ba- sis. Note that because of the comparatively low star formation rates (and correspondingly low Hα fluxes) of 3.3. Quantifying Star Formation and Neutral Gas Properties the SHIELD galaxies, Teich et al. (2016) did not signif- icantly interpret Hα-based radial profiles and did not As discussed in the preceding sections, the widespread show the Hα-based pixel-by-pixel correlation functions. recent star formation in NGC 5238 is occurring in a vari- In contrast, in NGC5238 the higher star formation rate ety of physical conditions. The regions of highest H I and more widespread Hα emission allows us to exam- mass surface density are Hα and UV-luminous; how- ine trends using both instantaneous and longer-timescale ever, the H I column density maximum is not coincident metrics. with the UV or with the Hα maxima. Similarly, there is In order to extract these diagnostics, the FUV, Hα, and some recent star formation occurring in regions of lower H I moment 0 images are placed on the same coordinate H I column densities (i.e., outside of the lowest contour grid. For the radially averaged profiles, the HST F606W 21 −2 shown in Figure 5, NHI = 10 cm ); such emission is image is used to fit stellar surface brightness as a function especially prominent in the western portion of the disk. of radius with the CleanGalaxy code (Hagen et al. We now seek to quantify the degree of co-spatiality 2014). This program identifies a central aperture lo- (or lack thereof) between H I mass surface density and cation (in this case, spatially coincident with the FUV the tracers of recent star formation (FUV and Hα emis- and Hα surface brightness maxima), the major and mi- sion). Numerous recent works have investigated em- nor axis lengths of concentric ellipses in the fit, and the pirical thresholds for star formation in the neutral and position angle of those ellipses. For NGC 5238, the el- 7

Figure 5. Comparison of high mass surface density H I gas with multi-wavelength imaging of NGC 5238. H I column density contours are shown at levels of (10,12,14,16) × 1020 cm−2 by white, yellow, orange, and red contours, respectively. Panel (a) shows the 3-color HST image from the LEGUS survey (F275W in blue, F336W in green, F438W in red). Panel (b) shows the 3-color HST image using archival ACS imaging (F606W in blue, F814W in red, and the linear combination of the two filters as green). Panel (c) shows the UV emission detected by GALEX (far-UV channel in blue, near-UV channel in red, and the linear combination of the two filters as green). Panel (d) shows the continuum-subtracted Hα image from the LVL survey data products (Kennicutt et al. 2008). The arrow in panel (a) denotes the location of the Milky Way foreground star. lipses are centered at (α,δ)=13h34m42.7s, +51◦36′50.5′′ The results of the radially-averaged analysis are shown (J2000), with an axial ratio of 1.54 and a position angle in Figure 6. As expected based on the qualitatively sim- (measured east of north) of 160◦. Following the discus- ilar morphologies and the similar global star formation sion in Haurberg (2013), this corresponds to an optical rates in the FUV and in Hα, these radial profiles are are inclination of 50◦, which is used to deproject values per effectively the same using either tracer. The H I profile unit area. Surface brightnesses in H I, FUV, and Hα is less centrally peaked and follows a shallower slope as are then extracted using either these concentric aper- a function of increasing galactocentric radius. In gen- tures (for the radially-averaged method) or in individual, eral, this plot demonstrates that low H I mass surface matched pixels (for the pixel-by-pixel method). densities are typically associated with lower star forma- 8

2 α

1 0 ⊙

SFR ⊙ log 0 2 log

1 0 ⊙

00 SFR α

l 0 00 0 0 0 0 0

Figure 6. ΣHα SFR, ΣFUV SFR, and ΣH I and vs. radius for NGC 5238. These plots allow us to examine the radially averaged trends of star formation rate surface density and H I mass surface density. tion rate surface densities. However, it also makes clear ongoing and recent star formation in NGC5238 that is that the regions of highest star formation rate surface occurring outside the regions of densest H I gas. The density are associated with gas at a range of H I mass “canonical” 1021 cm−2 column density threshold (corre- −2 surface densities; in the inner ∼500 pc of the disk, the sponding to 7.9 M⊙ pc ) is denoted by a vertical dotted star formation rate surface densities vary by an order of line; while most of the highest star formation rate sur- magnitude in annuli where the H I mass surface density face densities occur above this H I mass surface density is effectively constant. level, this is not a requirement. A significant amount of The radial integration used to create Figure 6 smears star formation is associated with H I gas that is “sub- out localized variations in star formation rate surface critical”. density and H I mass surface density. To overcome this limitation, we quantify the localized relations between 4. THE DYNAMICS OF NGC 5238 H I, Hα, and FUV emission using the pixel-by-pixel cor- Low-mass galaxies such as NGC5238 offer a unique op- relation analysis shown in Figure 7. The top and bot- portunity to populate the low-mass, low-velocity end of α tom panels show the FUV and the H -based star forma- the baryonic Tully-Fisher relation (Tully & Fisher 1977; tion rate surface densities as functions of the H I mass see also McGaugh 2012, McNichols etal. 2016, and the surface density on a pixel-by-pixel basis. These plots various references therein). These low-mass systems are quantify the local star formation law via the Schmidt- critical for our understanding of this fundamental scal- Kennicutt formalism; the power-law index using each ing relation that holds over many orders of magnitude tracer is shown in each panel. The FUV slope (N = for more massive galaxies. Importantly, gas-rich systems ± α 1.46 0.02) is slightly steeper than the H slope (N = in this mass regime allow us to determine the rotational 1.17±0.01), suggesting a slightly stronger correlation be- dynamics by spectral lines; gas-poor systems must rely tween FUV-bright regions and the highest H I mass sur- on dispersion measurements from stars. face densities. However, these plots also highlight the As discussed at length in McNichols et al. (2016), ex- 9

±

− 2

⊙ −

−

− 1

⊙ −

α

± SFR −

− l −

−

− ⊙ l

Figure 7. ΣHα SFR vs. ΣH I (top) and ΣFUV SFR vs. ΣH I (bottom) for NGC 5238. All FUV and Hα pixel values higher than 5% of the maximum are plotted; for H I all pixels higher than 10% of the maximum are plotted. A positive slope indicates high FUV and H I emission in the same pixels. The dashed vertical line represents the column density threshold of 1×1021 atoms cm−2. The power-law best fit slopes are shown; these correspond to the Kennicutt-Schmidt indices discussed in the text. tracting unambiguous information about the dynamics H I line width and the complex kinematics apparent in of low-mass galaxies is notoriously difficult. That work the two-dimensional velocity field (Figure 4). identifies an empirical threshold of 15 km s−1, below For uniformity we thus perform a spatially resolved which pressure support and coherent rotation can no position-velocity (hereafter, P-V) analysis in NGC 5238. longer be differentiated. Modeling in three spatial di- We first identify the kinematic major axis by inspection mensions is preferable to using collapsed two-dimensional as that axis along which the projected velocity extent is representations of the data (compare, for example, the largest. This major axis P-V slice (of width equal to the channel maps shown in Figure 2 to the intensity weighted H I beam size, 18′′) is centered at α,δ = (13h34m42.52, velocity field shown in Figure 4). However, in most cases, +51◦36′49.0′′, J2000) and follows a position angle of 20◦ degeneracies between inclination and rotation velocity east of north; this axis is evident in the two-dimensional make three-dimensional work insufficiently constrained velocity field shown in Figure 4. Note that this kinematic for unambiguous solutions. major axis not aligned with the optical major axis iden- Following the methodologies in McNichols et al. tified in § 3.3 (160◦ east of north, or 20◦ west of north); (2016), which include both two- and three-dimensional the two axes differ by ∼40◦. Such misalignments of the modeling approaches, we are not able to produce optical and H I major axes are common in the SHIELD an unambiguous model of the rotational dynamics of galaxies. NGC 5238 with the present data. Attempts were made Minor axis P-V slices (position angle = 290◦ east of to model the system as an asymmetrically distributed north) are extracted at positions along the kinematic gas disk exhibiting coherent rotation with a strong warp. major axis, separated by the H I beam width (18′′). These challenges may be expected based on the narrow 10

Figure 8. Spatially resolved P-V analysis of NGC 5238; all slices are the width of the H I synthesized beam (18′′). The top panel shows the major axis slice along a position angle 20◦ east of north. The bottom nine panels show the minor axis slices, each separated from one another by 18′′; slice 0 passes through the middle of the major axis slice and through the H I mass surface density maximum. In all panels, the green line identifies the systemic velocity of the source; the red lines identify the maximal velocity extent to which rotation is evident; the yellow contours show (3, 5, 10) times the per-channel noise in the cube (1.4 mJy Bm−1). From this analysis we infer a projected rotation velocity of 24 km s−1, at a maximum angular extent of 60′′. Each minor axis slice has a width equal to the H I has a projected rotational velocity of 24 kms−1, and that beam size. In total, nine such minor axis slices are this rotation can be unambiguously identified to a max- extracted; the central minor axis slice passes through imum physical radius of 1.3 kpc. the center of the major axis slice. These minor axis As discussed at length in McNichols et al. (2016), the slices show the intensity-weighted H I velocity and dis- correction of the observed velocities for inclination is the persion at a given position in the disk. As demonstrated most significant uncertainty in the analysis of the dy- in Cannon et al. (2011a), Bernstein-Cooper et al. (2014), namics of low-mass galaxies. Following that work, we and McNichols et al. (2016), if projected rotation is use the optical inclination derived in § 3.3 to correct the present in a system, then the minor-axis slices will iden- observed H I for projection effects and thus to deter- tify any such gradient as a change in the velocity of the mine the true rotational velocity. Our inclination value H I centroid as a function of position of the slice. of (50±5)◦ compares favorably with the value of 55◦ im- In Figure 8 we present the results of this spatially re- plied by the axial ratios derived from the UV images in solved P-V analysis. The major axis P-V slice identifies Lee et al. (2011); our measurement agrees within ∼5◦ a total velocity gradient of 48 km s−1, spanning 60′′ (1.3 with the implied inclination of 40◦ from the Dale et al. kpc at the adopted distance of 4.51 Mpc). This is con- (2009) IR apertures. firmed in the minor axis slices, which show a similar total The resulting circular velocity of NGC 5238 is 31 ± 5 velocity gradient across the disk. Note that the asym- kms−1, measured to a maximum physical radius of 1.3 metries of the NGC5238 neutral gas disk result in some kpc using gas detected at 3σ significance or higher. The 2 turnover in the centroid velocities of the H I gas in the v R 8 implied dynamical mass using M = G is (3±1) × 10 most distant minor axis slices (numbers +4 and −4 in 8 ′′ M⊙. This can be compared to 3.7 × 10 M⊙ when Figure 8); these slices are 72 (1.57 kpc) from the slice also explicitly accounting for random motions in the center. From this analysis we conclude that NGC 5238 H I gas as demonstrated in (Hoffman et al. 1996); for 11

Figure 9. The Baryonic Tully-Fisher relation (BTFR) as derived in McNichols et al. (2016). The small points are drawn from McGaugh (2012); the purple circles correspond to spiral galaxies; the gold diamonds represent less massive gas-rich galaxies; the red squares represent spheroidal dwarf galaxies with no detectable H I. The cyan diamond represents Leo P, the slowest rotating and lowest-mass galaxy known to still be relatively rich with interstellar gas (Bernstein-Cooper et al. 2014). The blue squares represent the SHIELD galaxies. The large magenta circle is our new measurement for NGC 5238. The light and dark shaded gray regions represent the 1 σ and the 3 σ deviations from a fit of the BTFR to the gas-rich galaxy sample, respectively. Within measurement errors, NGC 5238 lies on this calibration of the BTFR. this calculation we assume the average H I velocity NGC 5238 placed it inside the ALFALFA survey foot- dispersion throughout the disk (8.5 km s−1; see Fig- print (Giovanelli et al. 2005), it would not have been ure 4). Correcting the total H I mass for Helium and not met all of the SHIELD survey selection criteria. It −1 other metals (a 35% correction), the total gas mass of would meet the W50 < 65 kms line width criterion 7 7.2 NGC 5238 is (3.7±0.3) × 10 M⊙. Using the stellar mass but would fail the H I mass criterion (log[MH I] < 10 ); 8 in Table 1, the total baryonic mass is 1.3 × 10 M⊙; NGC 5238 is ∼75% too H I-massive to have been selected NGC 5238 is dark-matter dominated at a roughly 2.4:1 for SHIELD. Note, however, that stellar-based distances ratio, a value like those found for the SHIELD galax- to many of the SHIELD galaxies (McQuinn et al. 2014) ies in McNichols et al. (2016). These values allow us moved them further than the initial distance estimates to place NGC 5238 on the baryonic Tully-Fisher rela- upon selection. The final SHIELD sample presented in tion (BTFR). As shown in Figure 9, NGC 5238 agrees Teich et al. (2016) and McNichols et al. (2016) contains 7.2 with the calibration of the BTFR presented in McGaugh 4 sources with log(MH I) > 10 ; one of these sources (2012) and updated to include the SHIELD sample in (AGC749237) is more H I-massive than NGC 5238. McNichols et al. (2016). NGC 5238 follows the same Qualitatively, the physical properties of NGC 5238 scaling relation between rotational velocity and baryonic bridge the regime between the extremely low-mass mass as the SHIELD galaxies. SHIELD galaxies and the more massive dwarf irregular galaxies in the Local Group and beyond. As Figure 9 5. CONTEXTUALIZING NGC 5238 demonstrates, this source populates the region of the NGC5238 shares many physical characteristics with BTFR that is sparesly populated by the highest-mass the most massive galaxies from the SHIELD sample. SHIELD galaxies and the lowest-mass galaxies studied I Formally, if the Right Ascension and Declination of in McGaugh (2012). The H disk of NGC 5238 is larger 12

(physical diameter ≃4 kpc at largest extent) than those meates much of the inner disk. The most luminous UV of all of the SHIELD galaxies except for AGC749237, clusters are associated with high column density gas (NHI which is of comparable size. The peak H I column density ≃ 1021 cm−2), but there are a significant number of UV- in NGC5238 is the same as the highest peaks seen in the bright stars in regions of lower column densities as well. SHIELD galaxies (AGC110482 and AGC749237). In- The regions of highest H I mass surface densities are not terestingly, the kinematics of NGC 5238 are significantly co-spatial with the most luminous UV and Hα sources. more confused than those of the most massive SHIELD We quantify the degree of co-spatiality between H I and galaxies; AGC749237, AGC112521, and AGC110482 star formation tracers using both radially averaged and show comparatively ordered rotation in two-dimensional pixel-by-pixel star formation rate density analyses. The velocity fields. NGC 5238, in contrast, shows complex far-UV results are qualitatively similar to those found for kinematics; two- and three- dimensional rotationally- the properties of star formation in the SHIELD galaxies supported disk modeling fails in this comparatively mas- by Teich et al. (2016). The Hα emission in NGC 5238 is sive source. significantly more widespread than in the SHIELD galax- As discussed in Section 3.2, the FUV and Hα star for- ies. The Kennicutt-Schmidt indices are N = 1.46±0.02 mation rates of NGC5238 are higher than those of any and N = 1.17±0.01 in the FUV and in Hα, respectively. of the SHIELD galaxies (Teich et al. 2016). The origin of The Hα-based value is steeper than the value found the high star formation rate in NGC5238 compared to for the composite SHIELD sample (N = 0.68±0.04) in those in the SHIELD galaxies is not immediately clear. Teich et al. (2016), but is in good agreement with the There are no known galaxies cataloged in the NASA values found in the individual SHIELD galaxies with the Extragalactic Database8 within a projected separation highest star formation rates and the highest H I mass of 300 kpc whose recessional velocities are within 100 surface densities. −1 −1 kms of the VHI = 232 ± 1 kms derived in this work The complex neutral gas dynamics of NGC5238 pre- (see discussions in Lelli et al. 2014 and Martinkus et al. clude an unambiguous two or three-dimensional model of 2015). While the outer H I disk of NGC5238 does show the rotation of the source; degeneracies persist between some asymmetries, we do not detect unambiguous evi- inclination and rotation velocity. As for the SHIELD dence for an ongoing tidal interaction at the current sen- galaxies analyzed in McNichols et al. (2016), we use sitivity and angular resolution of our VLA observations. spatially-resolved position velocity analysis to estimate We note that recent work suggests that interactions at the projected rotation velocity of the galaxy. We cor- large distances may not be important for star formation rect this value for inclination based on an elliptical fit rates in dwarf galaxies (see, e.g., Pearson et al. 2016). to the red stellar population seen by HST. The resulting −1 The distributed recent star formation in NGC 5238 dif- rotational velocity Vrot = (31 ± 5) kms implies a total 8 ferentiates it from most of the SHIELD galaxies. In dynamical mass of 3 × 10 M⊙. Comparing to the sum those systems, almost all of the galaxies show Hα emis- of the neutral gas and stellar , NGC5238 is dark sion that is concentrated to a few isolated HII regions matter dominated at a ratio of roughly 2.4:1. The galaxy or at most some faint diffuse emission. In contrast, falls on the baryonic Tully-Fisher relation as presented NGC5238 harbors widespread Hα emission with both in McNichols et al. (2016). clustered and diffuse morphology. From the far-UV per- NGC 5238 is an intriguing system that warrants further spective, the SHIELD galaxies show a variety of mor- study. Higher angular resolution H I imaging could re- phologies; some sources have widespread far-UV emis- veal small-scale feedback processes between the multiple sion throughout most of the inner disk, while others show stellar clusters and the surrounding neutral gas. The stel- only very faint far-UV emission. Again NGC 5238 stands lar population is well-resolved by HST, making possible out in a comparative sense: the entire inner disk is UV- a detailed study of the recent star formation history. The bright, which is unlike the SHIELD galaxies. significant nebular emission would facilitate detailed and 6. CONCLUSIONS resolved analysis of the chemical composition of the sys- tem. The preliminary estimates of the metallicity (Z ≃ We have presented new VLA H I spectral line observa- 19%Z⊙; Moustakas & Kennicutt 2006) place this system tions of the nearby star-forming dwarf near the empirical boundary below which CO emission NGC 5238. The data resolve the structure and dynamics becomes extremely difficult to detect in dwarf galaxies of the neutral gas on spatial (∼400 pc) and spectral (1.56 − − (e.g., Taylor et al. 1998; Leroy et al. 2005; Schruba et al. kms 1 ch 1) scales that can be meaningfully compared 2012). While single-dish CO observations by Leroy et al. to other major H I survey results. The outer portions of (2005) failed to detect NGC 5238, the high star forma- the neutral gas disk show significant asymmetries. The tion rate and significant H I mass surface densities sug- high mass surface density H I gas displays an interest- gest that targeted interferometric observations could be ing crescent-shaped morphology, with the highest H I fruitful. column densities in excess of 1.5 × 1021 cm−2. The new H I images are compared with images from HST, GALEX, and the Bok 2.3m telescope in order The authors are immensely grateful to the National to study the nature of the recent star formation in Radio Astronomy Observatory for making the “Observ- NGC 5238. The stellar disk is resolved into multiple stel- ing for University Classes” program available to the as- lar clusters by the HST LEGUS images (Calzetti et al. tronomical and teaching communities. We would like to 2015), and there is an exquisite match between the UV- personally thank Vivek Dhawan, Miller Goss, Christo- bright stellar population and the Hα emission that per- pher Hales, Amy Kimball, Chris Langley, Emmanuel

8 Momjian, J¨urgen Ott, Lorant Sjouwerman, and Gustaaf http://ned.ipac.caltech.edu van Moorsel for making our visit to the Domenici Science 13

Operations Center and the VLA site productive and en- Dreyer, J. L. E. 1888, MmRAS, 49, 1 joyable. We are grateful to Macalester College for sup- Elmegreen, B. G., & Hunter, D. A. 2015, ApJ, 805, 145 porting this project. We would like to thank the referee Giovanelli, R., Haynes, M. P., Kent, B. R., et al. 2005, AJ, 130, for a prompt and insightful report that helped to improve 2598 Hagen, C., Cannon, J. M., Cave, I., et al. 2014, American this manuscript. Astronomical Society Meeting Abstracts #223, 223, #355.16 Funding for the Sloan Digital Sky Survey IV has been Hao, C.-N., Kennicutt, R. C., Johnson, B. D., et al. 2011, ApJ, provided by the Alfred P. Sloan Foundation, the U.S. 741, 124 Department of Energy Office of Science, and the Partici- Haurberg, N. C. 2013, Ph.D. Thesis, Indiana University pating Institutions. SDSS-IV acknowledges support and Hoffman, G. L., Salpeter, E. E., Farhat, B., et al. 1996, ApJS, 105, 269 resources from the Center for High-Performance Com- Hunter, D. A., Elmegreen, B. G., & Baker, A. L. 1998, ApJ, 493, puting at the University of Utah. The SDSS web site is 595 www.sdss.org. Hunter, D. A., Ficut-Vicas, D., Ashley, T., et al. 2012, AJ, 144, SDSS-IV is managed by the Astrophysical Research 134 Ianjamasimanana, R., de Blok, W. J. G., Walter, F., et al. 2015, Consortium for the Participating Institutions of the AJ, 150, 47 SDSS Collaboration including the Brazilian Partici- Jorsater, S., & van Moorsel, G. A. 1995, AJ, 110, 2037 pation Group, the Carnegie Institution for Science, Kennicutt, R. C., Jr., Armus, L., Bendo, G., et al. 2003, PASP, Carnegie Mellon University, the Chilean Participa- 115, 928 tion Group, the French Participation Group, Harvard- Kennicutt, R. C., Jr., Lee, J. C., Funes, J. G., et al. 2008, ApJS, 178, 247-279 Smithsonian Center for Astrophysics, Instituto de As- Kennicutt, R. C., Jr. 1998, ARA&A, 36, 189 trof´ısica de Canarias, The Johns Hopkins University, Kennicutt, R. C., Calzetti, D., Aniano, G., et al. 2011, PASP, Kavli Institute for the Physics and Mathematics of 123, 1347 the Universe (IPMU) / University of Tokyo, Lawrence Kirby, E. M., Koribalski, B., Jerjen, H., & L´opez-S´anchez, A.´ Berkeley National Laboratory, Leibniz Institut f¨ur As- 2012, MNRAS, 420, 2924 trophysik Potsdam (AIP), Max-Planck-Institut f¨ur As- Lee, J. C., Gil de Paz, A., Kennicutt, R. C., Jr., et al. 2011, ApJS, 192, 6 tronomie (MPIA Heidelberg), Max-Planck-Institut f¨ur Lelli, F., Verheijen, M., & Fraternali, F. 2014, MNRAS, 445, 1694 Astrophysik (MPA Garching), Max-Planck-Institut f¨ur Leroy, A., Bolatto, A. D., Simon, J. D., & Blitz, L. 2005, ApJ, Extraterrestrische Physik (MPE), National Astronomi- 625, 763 cal Observatory of China, New Mexico State University, Marble, A. R., Engelbracht, C. W., van Zee, L., et al. 2010, ApJ, New York University, University of Notre Dame, Obser- 715, 506 Martinkus, C., Cannon, J. M., McQuinn, K. B., et al. 2015, vat´ario Nacional / MCTI, The Ohio State University, American Astronomical Society Meeting Abstracts, 225, 248.21 Pennsylvania State University, Shanghai Astronomical McGaugh, S. S. 2012, AJ, 143, 40 Observatory, United Kingdom Participation Group, Uni- McNichols, A. T., Teich, Y. G., Nims, E., Cannon, J. M., et al. versidad Nacional Aut´onoma de M´exico, University of 2016, ApJ, in press (ArXiV/1609.05376) Arizona, University of Colorado Boulder, University of McQuinn, K. B. W., Cannon, J. M., Dolphin, A. E., et al. 2014, ApJ, 785, 3 Oxford, University of Portsmouth, University of Utah, McQuinn, K. B. W., Skillman, E. D., Dolphin, A. E., & Mitchell, University of Virginia, University of Washington, Univer- N. P. 2015, ApJ, 808, 109 sity of Wisconsin, Vanderbilt University, and Yale Uni- Mogotsi, K. M., de Blok, W. J. G., Cald´u-Primo, A., et al. 2016, versity. AJ, 151, 15 Moustakas, J., & Kennicutt, R. C., Jr. 2006, ApJS, 164, 81 Facilities: VLA, HST, GALEX, Bok 2.3m Ott, J., Stilp, A. M., Warren, S. R., et al. 2012, AJ, 144, 123 Pearson, S., Besla, G., Putman, M. E., et al. 2016, MNRAS, 459, REFERENCES 1827 Roychowdhury, S., Chengalur, J. N., Kaisin, S. S., & Asplund, M., Grevesse, N., Sauval, A. J., & Scott, P. 2009, Karachentsev, I. D. 2014, MNRAS, 445, 1392 ARA&A, 47, 481 Schmidt, M. 1959, ApJ, 129, 243 Begum, A., Chengalur, J. N., Karachentsev, I. D., Sharina, M. E., Schruba, A., Leroy, A. K., Walter, F., et al. 2012, AJ, 143, 138 & Kaisin, S. S. 2008, MNRAS, 386, 1667 Skillman, E. D. 1987, NASA Conference Publication, 2466, 263 Bernstein-Cooper, E. Z., Cannon, J. M., Elson, E. C., et al. 2014, Swaters, R. A., van Albada, T. S., van der Hulst, J. M., & AJ, 148, 35 Sancisi, R. 2002, A&A, 390, 829 Bigiel, F., Leroy, A., Walter, F., et al. 2008, AJ, 136, 2846 Tamburro, D., Rix, H.-W., Leroy, A. K., et al. 2009, AJ, 137, 4424 Calzetti, D., Lee, J. C., Sabbi, E., et al. 2015, AJ, 149, 51 Taylor, C. L., Kobulnicky, H. A., & Skillman, E. D. 1998, AJ, Cannon, J. M., Most, H. P., Skillman, E. D., et al. 2011, ApJ, 116, 2746 735, 35 Teich, Y. G., McNichols, A.T., Nims, E., Cannon, J. .M., et al. Cannon, J. M., Giovanelli, R., Haynes, M. P., et al. 2011, ApJ, 2016, ApJ, in press (ArXiV/1609.05375) 739, L22 Thuan, T. X., Lipovetsky, V. A., Martin, J.-M., & Pustilnik, Cannon, J. M., Martinkus, C. P., Leisman, L., et al. 2015, AJ, S. A. 1999, A&AS, 139, 1 149, 72 Tully, R. B., & Fisher, J. R. 1977, A&A, 54, 661 Cook, D. O., Dale, D. A., Johnson, B. D., et al. 2014, MNRAS, Tully, R. B., Rizzi, L., Shaya, E. J., et al. 2009, AJ, 138, 323 445, 881 van Zee, L., Haynes, M. P., Salzer, J. J., & Broeils, A. H. 1997, Cook, D. O., Dale, D. A., Johnson, B. D., et al. 2014, MNRAS, AJ, 113, 1618 445, 899 York, D. G., Adelman, J., Anderson, J. E., Jr., et al. 2000, AJ, Dalcanton, J. J., Williams, B. F., Seth, A. C., et al. 2009, ApJS, 120, 1579 183, 67 Dale, D. A., Cohen, S. A., Johnson, L. C., et al. 2009, ApJ, 703, 517